top research papers of 2022

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

• View all journals
• Explore content
• About the journal
• Publish with us
• Sign up for alerts

Collection  07 April 2023

Top 25 Life and Biological Sciences Articles of 2022

We are pleased to share with you the 25 most downloaded Nature Communications articles* in the life and biological sciences published in 2022. (Please note we have a separate collection on the Top 25 COVID-19 papers .) Featuring authors from around the world, these papers highlight valuable research from an international community. 

Browse all Top 25 subject area collections .

*Data obtained from SN Insights (based on Digital Science's Dimensions) and have been normalised to account for articles published later in the year.

Top 25 Articles

Engineered helicase replaces thermocycler in DNA amplification while retaining desired PCR characteristics

PCR is an essential method for the amplification and manipulation of nucleic acids, but the requirement for a thermocycler limits access. Here, authors engineer a helicase to replace the heating step of PCR with enzymatic unwinding, allowing the isothermal amplification of fragments up to 6 kb.

• Momčilo Gavrilov
• Joshua Y. C. Yang

Evolution of sexual systems, sex chromosomes and sex-linked gene transcription in flatworms and roundworms

Transitions between hermaphroditic and separate sexes are relatively understudied in animals compared to pants. Here, Wang et al. reconstruct the evolution of separate sexes in the flatworms and complex changes of sex chromosomes in the roundworms.

• Yifeng Wang
• Robin B. Gasser

The gut microbiota and depressive symptoms across ethnic groups

Here, by studying a multi-ethnic cross-sectional urban cohort ( N  = 3211, 6 ethnic groups), the authors show that depressive symptom levels are related to the gut microbiota taxonomic characteristics but that these are largely invariant across ethnic groups.

• Jos A. Bosch
• Max Nieuwdorp

Metagenomics of Parkinson’s disease implicates the gut microbiome in multiple disease mechanisms

Here, the authors perform large-scale high-resolution Parkinson’s disease metagenomics analyses, revealing widespread dysbiosis characterized by overabundance of pathogens, immunogens, toxicants, and curli, reduction in neuroprotective and antiinflammatory molecules, and dysregulated neuroactive signaling.

• Zachary D. Wallen
• Ayse Demirkan
• Haydeh Payami

Codon-specific Ramachandran plots show amino acid backbone conformation depends on identity of the translated codon

Genetic code redundancies are considered inconsequential to protein structure. This study uncovers a dependence between local amino acid conformation in folded proteins and the identity of the codon from which that amino acid was translated.

• Aviv A. Rosenberg
• Alex M. Bronstein

De novo analysis of bulk RNA-seq data at spatially resolved single-cell resolution

Current methods to reanalyze bulk RNA-seq at spatially resolved single-cell resolution have limitations. Here, the authors develop Bulk2Space, a spatial deconvolution algorithm using single-cell and spatial transcriptomics as references, providing new insights into spatial heterogeneity within bulk tissue.

• Jingyang Qian
• Xiaohui Fan

Exceptional preservation and foot structure reveal ecological transitions and lifestyles of early theropod flyers

The shape of bird toe pads and foot scales can be used to infer their behaviour. Here, the authors examine fossil evidence of toe pads and scales, in addition to claws and bones, from birds and close relatives, illustrating diverse lifestyles and ecological roles among early theropod flyers.

• Michael Pittman
• Phil R. Bell
• Thomas G. Kaye

Common evolutionary origin of acoustic communication in choanate vertebrates

Here, the authors record acoustic communication in 53 species commonly considered non-vocal and reconstruct acoustic communication as originating 407 million years ago.

• Gabriel Jorgewich-Cohen
• Simon William Townsend
• Marcelo R. Sánchez-Villagra

Calcium-mediated rapid movements defend against herbivorous insects in Mimosa pudica

Mimosa pudica moves its leaves within seconds of being touched or wounded. Here the authors show that such movements are triggered by rapid changes in Ca 2+ and action and variation potentials and provide evidence that rapid movements help protect the plant from insect attacks.

• Takuma Hagihara
• Hiroaki Mano
• Masatsugu Toyota

Evolutionary transition from a single RNA replicator to a multiple replicator network

Long-term experimental evolution shows that a single polymerase-encoding RNA replicator can evolve into a complex replicator network, shedding light on how a molecular replicator could have developed complexity before the emergence of life.

• Ryo Mizuuchi
• Taro Furubayashi
• Norikazu Ichihashi

Lysosomal damage drives mitochondrial proteome remodelling and reprograms macrophage immunometabolism

Extensive lysosomal damage can result in cell death but how limited protease leakage affects cytoplasmic organelles in viable cells is not well understood. Here the authors show that limited lysosomal damage leads to changes in the mitochondrial proteome and the modulation of macrophage immunometabolism.

• Claudio Bussi
• Tiaan Heunis
• Maximiliano G. Gutierrez

Payload distribution and capacity of mRNA lipid nanoparticles

Lipid nanoparticles (LNPs) are effective vehicles to deliver mRNA vaccines and therapeutics but assessing the mRNA packaging characteristics in LNPs is challenging. Here, the authors report that mRNA and lipid contents in LNP formulations can be quantitatively examined by multi-laser cylindrical illumination confocal spectroscopy at the single-nanoparticle level.

• Hai-Quan Mao

Improved prediction of protein-protein interactions using AlphaFold2

Predicting the structure of protein complexes is extremely difficult. Here, authors apply AlphaFold2 with optimized multiple sequence alignments to model complexes of interacting proteins, enabling prediction of both if and how proteins interact with state-of-art accuracy.

• Patrick Bryant
• Gabriele Pozzati
• Arne Elofsson

Predicting the structure of large protein complexes using AlphaFold and Monte Carlo tree search

The accuracy of AlphaFold decreases with the number of protein chains and the available GPU memory limits the size of protein complexes that can be predicted. Here, the authors show that complexes with 10–30 chains can be assembled from predicted subcomponents using Monte Carlo tree search.

Microbiome differential abundance methods produce different results across 38 datasets

Many microbiome differential abundance methods are available, but it lacks systematic comparison among them. Here, the authors compare the performance of 14 differential abundance testing methods on 38 16S rRNA gene datasets with two sample groups, and show ALDEx2 and ANCOM-II produce the most consistent results.

• Jacob T. Nearing
• Gavin M. Douglas
• Morgan G. I. Langille

Deep phenotyping and lifetime trajectories reveal limited effects of longevity regulators on the aging process in C57BL/6J mice

Lifespan can be affected by both physiological ageing and specific sets of pathologies associated with old age. Here the authors report a resource of large-scale cross-sectional phenotyping of aging male mice at different time points to analyse a large set of phenotypes and molecular markers, including during genetic and diet interventions affecting lifespan.

• Helmut Fuchs
• Dan Ehninger

Structure-based discovery of small molecules that disaggregate Alzheimer’s disease tissue derived tau fibrils in vitro

Evidence suggests that fibrous aggregates of protein tau may be the proximal cause of Alzheimer’s disease. Here, using atomic structures of tau fibrils from brains of people with Alzheimer’s disease, the authors have found small-molecule drug leads that disaggregate tau fibrils in vitro.

• Paul M. Seidler
• Kevin A. Murray
• David S. Eisenberg

The intestinal clock drives the microbiome to maintain gastrointestinal homeostasis

Here, Heddes et al. demonstrate a major role for the intestinal circadian clock in driving microbiome dynamics. Microbiota transfer from intestinal clock-deficient mice promotes altered intestinal phenotypes, highlighting the importance of functional intestinal clocks for gastrointestinal homeostasis of the host.

• Marjolein Heddes
• Baraa Altaha
• Silke Kiessling

A new gene set identifies senescent cells and predicts senescence-associated pathways across tissues

Identification of senescent cells in vivo remains a challenging task. Here the authors present and validate a senescence gene set called SenMayo enriched in human and murine aged tissues.

• Dominik Saul
• Robyn Laura Kosinsky
• Sundeep Khosla

Caffeine blocks SREBP2-induced hepatic PCSK9 expression to enhance LDLR-mediated cholesterol clearance

Caffeine may reduce cardiovascular disease risk, but the underlying mechanisms for these effects are incompletely understood. Here the authors report that caffeine inhibits the activation of the transcription factor SREBP2 to promote LDLc clearance through the PCSK9-LDLR axis.

• Paul F. Lebeau
• Jae Hyun Byun
• Richard C. Austin

Tissue-specific impacts of aging and genetics on gene expression patterns in humans

Age is a risk factor for many diseases, but the impact of aging on molecular phenotypes is not fully understood. Here, the authors quantify the relative contributions of genetics and aging to gene expression patterns across 27 tissues in humans, showing that age and genetics each play distinct roles in shaping expression phenotypes.

• Ryo Yamamoto
• Peter H. Sudmant

Altered glycolysis triggers impaired mitochondrial metabolism and mTORC1 activation in diabetic β-cells

Chronic hyperglycemia impairs insulin secretion from pancreatic beta cells in diabetes. Here, the authors reveal that a glucose metabolite is responsible and show lowering glucose metabolism during hyperglycemia prevents loss of beta-cell function.

• Elizabeth Haythorne
• Matthew Lloyd
• Frances M. Ashcroft

Perspectives in machine learning for wildlife conservation

Animal ecologists are increasingly limited by constraints in data processing. Here, Tuia and colleagues discuss how collaboration between ecologists and data scientists can harness machine learning to capitalize on the data generated from technological advances and lead to novel modeling approaches.

• Benjamin Kellenberger
• Tanya Berger-Wolf

Fully-automated and ultra-fast cell-type identification using specific marker combinations from single-cell transcriptomic data

Cell types are typically identified in single cell transcriptomic data by manual annotation of cell clusters using established marker genes. Here the authors present a fully-automated computational platform that can quickly and accurately distinguish between cell types.

• Aleksandr Ianevski
• Anil K. Giri
• Tero Aittokallio

Host-dependent resistance of Group A Streptococcus to sulfamethoxazole mediated by a horizontally-acquired reduced folate transporter

There is increasing evidence for metabolic processes mediating antimicrobial resistance. Here, the authors present a mechanism of sulfamethoxazole resistance in Group A Streptococcus that is dependent on acquiring end products of the host folate biosynthesis pathway.

• M. Kalindu D. Rodrigo
• Aarti Saiganesh
• Timothy C. Barnett

Quick links

• Explore articles by subject
• Guide to authors
• Editorial policies

• SUGGESTED TOPICS
• The Magazine
• Newsletters
• Managing Yourself
• Managing Teams
• Work-life Balance
• The Big Idea
• Data & Visuals
• Reading Lists
• Case Selections
• HBR Learning
• Topic Feeds
• Account Settings
• Email Preferences

HBR’s Most-Read Research Articles of 2022

• Dagny Dukach

Insights on equity, leadership, and becoming your best self.

The new year is a great time to set ambitious goals. But alongside our plans for the future, it’s also helpful to acknowledge all the challenges we’ve faced — and the progress we’ve made — in the last 12 months. In this end-of-year roundup, we share key insights and trends from HBR’s most-read research articles of 2022, exploring topics from embracing a new identity to fostering equity in the workplace and beyond.

For many of us, the arrival of a new year can be equal parts inspiring and daunting. While the promise of a fresh start is often welcome, it’s also a reminder of all the challenges we faced in the last 12 months — and all those still awaiting us, that we have yet to overcome.

• Dagny Dukach is a former associate editor at Harvard Business Review.

Partner Center

Frontiers | Science News

• Science News

Research Topics

Top 10 research topics from 2022.

Find the answers to your biggest research questions from 2022. With collective views of over 3.2 million, researchers explored topics spanning from vaccine safety and psychedelic therapy to quaternary fossils and antiviral plants .

Research Topics:

1. Viral diseases

• 499,000 views
• 37 articles

2. Exploring sound

• 442,000 views
• 23 articles

3. Neurodegenerative diseases

• 387,000 views
• 19 articles

4. Psychedelic therapy

• 346,000 views
• 22 articles

5. Circadian rhythms

• 335,000 views

6. Covid-19 insights

• 320,000 views
• 84 articles

7. Vaccine safety

• 244,000 views
• 10 articles

8. Spirituality and healthcare

• 242,000 views
• 16 articles

9. Antiviral plants

• 208,000 views

10. Quaternary fossils

• 144,000 views

Shape the future of your field

Become a guest editor for an article collection around your own research theme. Benefit from increased impact and discoverability, a dedicated platform and support team, and rigorous peer review for every paper.

Suggest your topic

Post related info

December 09, 2022

Frontiers Communications

Post categories, related subjects, research topics, related content.

Frontiers Research Topics has a new look

Second annual Frontiers Spotlight Award announced for 2018

What is a Frontiers Research Topic, and why should I participate in one?

Latest posts.

AI, open science, and the extreme weather pandemic: Takeaways from SXSW 2024

Big data, AI, and personalized medicine: scientists reveal playbook aiming to revolutionize healthcare

Registration open: Dr Eric Topol to explore how AI will shape the future of healthcare at Frontiers Forum virtual event

Babies in the womb exposed to two languages hear speech differently when born

Screen time not the main factor making parent-child interactions worse, study finds

• Event calendar

The Top 17 ‘Must-Read’ AI Papers in 2022

We caught up with experts in the RE•WORK community to find out what the top 17 AI papers are for 2022 so far that you can add to your Summer must reads. The papers cover a wide range of topics including AI in social media and how AI can benefit humanity and are free to access.

Interested in learning more? Check out all the upcoming RE•WORK events to find out about the latest trends and industry updates in AI here .

Max Li, Staff Data Scientist – Tech Lead at Wish

Max is a Staff Data Scientist at Wish where he focuses on experimentation (A/B testing) and machine learning.  His passion is to empower data-driven decision-making through the rigorous use of data. View Max’s presentation, ‘Assign Experiment Variants at Scale in A/B Tests’, from our Deep Learning Summit in February 2022 here .

1. Boostrapped Meta-Learning (2022) – Sebastian Flennerhag et al.

The first paper selected by Max proposes an algorithm in which allows the meta-learner teach itself, allowing to overcome the meta-optimisation challenge. The algorithm focuses meta-learning with gradients, which guarantees improvements in performance. The paper also looks at how bootstrapping opens up possibilities. Read the full paper here .

2. Multi-Objective Bayesian Optimization over High-Dimensional Search Spaces (2022) – Samuel Daulton et al.

Another paper selected by Max proposes MORBO, a scalable method for multiple-objective BO as it performs better than that of high-dimensional search spaces. MORBO significantly improves the sample efficiency, and where BO algorithms fail, MORBO provides improved sample efficiencies to the current BO approach used. Read the full paper here .

3. Tabular Data: Deep Learning is Not All You Need (2021) – Ravid Shwartz-Ziv, Amitai Armon

To solve real-life data science problems, selecting the right model to use is crucial. This final paper selected by Max explores whether deep models should be recommended as an option for tabular data. Read the full paper here .

Jigyasa Grover, Senior Machine Learning Engineer at Twitter

Jigyasa Grover is a Senior Machine Learning Engineer at Twitter working in the performance ads ranking domain. Recently, she was honoured with the 'Outstanding in AI: Young Role Model Award' by Women in AI across North America. She is one of the few ML Google Developer Experts globally. Jigyasa has previously presented at our Deep Learning Summit and MLOps event in San Fransisco earlier this year.

4. Privacy for Free: How does Dataset Condensation Help Privacy? (2022) – Tian Dong et al.

Jigyasa’s first recommendation concentrates on Privacy Preserving Machine Learning, specifically mitigating the leakage of sensitive data in Machine Learning. The paper provides one of the first propositions of using dataset condensation techniques to preserve the data efficiency during model training and furnish membership privacy. This paper was published by Sony AI and won the Outstanding Paper Award at ICML 2022. Read the full paper here .

5. Affective Signals in a Social Media Recommender System (2022) – Jane Dwivedi-Yu et al.

The second paper recommended by Jigyasa talks about operationalising Affective Computing, also known as Emotional AI, for an improved personalised feed on social media. The paper discusses the design of an affective taxonomy customised to user needs on social media. It further lays out the curation of suitable training data by combining engagement data and data from a human-labelling task to enable the identification of the affective response a user might exhibit for a particular post. Read the full paper here .

6. ItemSage: Learning Product Embeddings for Shopping Recommendations at Pinterest (2022) – Paul Baltescu et al.

Jigyasa’s last recommendation is a paper by Pinterest that illustrates the aggregation of both textual and visual information to build a unified set of product embeddings to enhance recommendation results on e-commerce websites. By applying multi-task learning, the proposed embeddings can optimise for multiple engagement types and ensures that the shopping recommendation stack is efficient with respect to all objectives. Read the full article here .

Asmita Poddar, Software Development Engineer at Amazon Alexa

Asmita is a Software Development Engineer at Amazon Alexa, where she works on developing and productionising natural language processing and speech models. Asmita also has prior experience in applying machine learning in diverse domains. Asmita will be presenting at our London AI Summit , in September, where she will discuss AI for Spoken Communication.

7. Competition-Level Code Generation with AlphaCode (2022) – Yujia Li et al.

Systems can help programmers become more productive. Asmita has selected this paper which addresses the problems with incorporating innovations in AI into these systems. AlphaCode is a system that creates solutions for problems that requires deeper reasoning. Read the full paper here .

8. A Commonsense Knowledge Enhanced Network with Retrospective Loss for Emotion Recognition in Spoken Dialog (2022) – Yunhe Xie et al.

There are limits to model’s reasoning in regards to the existing ERSD datasets. The final paper selected by Asmita proposes a Commonsense Knowledge Enhanced Network with a backward-looking loss to perform dialog modelling, external knowledge integration and historical state retrospect. The model used has been shown to outperform other models. Read the full paper here .

Discover the speakers we have lined up and the topics we will cover at the London AI Summit.

Sergei Bobrovskyi, Expert in Anomaly Detection for Root Cause Analysis at Airbus

Dr. Sergei Bobrovskyi is a Data Scientist within the Analytics Accelerator team of the Airbus Digital Transformation Office. His work focuses on applications of AI for anomaly detection in time series, spanning various use-cases across Airbus. Sergei will be presenting at our Berlin AI Summit in October about Anomaly Detection, Root Cause Analysis and Explainability.

9. LaMDA: Language Models for Dialog Applications (2022) – Romal Thoppilan et al.

The paper chosen by Sergei describes the LaMDA system, which caused the furor this summer, when a former Google engineer claimed it has shown signs of being sentient. LaMDA is a family of large language models for dialog applications based on Transformer architecture. The interesting feature of the model is their fine-tuning with human annotated data and possibility to consult external sources. In any case, this is a very interesting model family, which we might encounter in many of the applications we use daily. Read the full paper here .

10. A Path Towards Autonomous Machine Intelligence Version 0.9.2, 2022-06-27 (2022) – Yann LeCun

The second paper chosen by Sergei provides a vision on how to progress towards general AI. The study combines a number of concepts including configurable predictive world model, behaviour driven through intrinsic motivation, and hierarchical joint embedding architectures. Read the full paper here .

11. Coordination Among Neural Modules Through a Shared Global Workpace (2022) – Anirudh Goyal et al.

This paper chosen by Sergei combines the Transformer architecture underlying most of the recent successes of deep learning with ideas from the Global Workspace Theory from cognitive sciences. This is an interesting read to broaden the understanding of why certain model architectures perform well and in which direction we might go in the future to further improve performance on challenging tasks. Read the full paper here .

12. Magnetic control of tokamak plasmas through deep reinforcement learning (2022) – Jonas Degrave et al.

Sergei chose the next paper, which asks the question of ‘how can AI research benefit humanity?’. The use of AI to enable safe, reliable and scalable deployment of fusion energy could contribute to the solution of pression problems of climate change. Sergei has said that this is an extremely interesting application of AI technology for engineering. Read the full paper here .

13. TranAd: Deep Transformer Networks for Anomaly Detection in Multivariate Time Series Data (2022) – Shreshth Tuli, Giuliano Casale and Nicholas R. Jennings

The final paper chosen by Sergei is a specialised paper applying transformer architecture to the problem of unsupervised anomaly detection in multivariate time-series. Many architectures which were successful in other fields are at some points being also applied to time-series. The paper shows an improved performance on some known data sets. Read the full paper here .

Abdullahi Adamu, Senior Software Engineer at Sony

Abdullahi has worked in various industries including working at a market research start-up where he developed models that could extract insights from human conversations about products or services. He moved to Publicis, where he became Data Engineer and Data Scientist in 2018. Abdullahi will be part of our panel discussion at the London AI Summit in September, where he will discuss Harnessing the Power of Deep Learning.

14. Self-Supervision for Learning from the Bottom Up (2022) – Alexei Efros

This paper chosen by Abdullahi makes compelling arguments for why self-supervision is the next step in the evolution of AI/ML for building more robust models. Overall, these compelling arguments justify even further why self-supervised learning is important on our journey towards more robust models that generalise better in the wild. Read the full paper here .

15. Neural Architecture Search Survey: A Hardware Perspective (2022) – Krishna Teja Chitty-Venkata and Arun K. Somani

Another paper chosen by Abdullahi understands that as we move towards edge computing and federated learning, neural architecture search that takes into account hardware constraints which will be more critical in ensuring that we have leaner neural network models that balance latency and generalisation performance. This survey gives a birds eye view of the various neural architecture search algorithms that take into account hardware constraints to design artificial neural networks that give the best tradeoff of performance and accuracy. Read the full paper here .

16. What Should Not Be Contrastive In Contrastive Learning (2021) – Tete Xiao et al.

In the paper chosen by Abdullahi highlights the underlying assumptions behind data augmentation methods and how these can be counter productive in the context of contrastive learning; for example colour augmentation whilst a downstream task is meant to differentiate colours of objects. The result reported show promising results in the wild. Overall, it presents an elegant solution to using data augmentation for contrastive learning. Read the full paper here .

17. Why do tree-based models still outperform deep learning on tabular data? (2022) – Leo Grinsztajn, Edouard Oyallon and Gael Varoquaux

The final paper selected by Abdulliah works on answering the question of why deep learning models still find it hard to compete on tabular data compared to tree-based models. It is shown that MLP-like architectures are more sensitive to uninformative features in data, compared to their tree-based counterparts. Read the full paper here .

Sign up to the RE•WORK monthly newsletter for the latest AI news, trends and events.

Join us at our upcoming events this year:

·       London AI Summit – 14-15 September 2022

·       Berlin AI Summit – 4-5 October 2022

·       AI in Healthcare Summit Boston – 13-14 October 2022

·       Sydney Deep Learning and Enterprise AI Summits – 17-18 October 2022

·       MLOps Summit – 9-10 November 2022

·       Toronto AI Summit – 9-10 November 2022

·       Nordics AI Summit - 7-8 December 2022

• Mar 2, 2023

Must read: the 100 most cited AI papers in 2022

Updated: Mar 8, 2023

Who Is publishing the most Impactful AI research right now? With the breakneck pace of innovation in AI, it is crucial to pick up some signal as soon as possible. No one has the time to read everything, but these 100 papers are sure to bend the road as to where our AI technology is going. The real test of impact of R&D teams is of course how the technology appears in products, and OpenAI shook the world by releasing ChatGPT at the end of November 2022, following fast on their March 2022 paper “Training language models to follow instructions with human feedback”. Such fast product adoption is rare, so to see a bit further, we look at a classic academic metric: the number of citations. A detailed analysis of the 100 most cited papers per year, for 2022, 2021, and 2020 allows us to draw some early conclusions. The United States and Google still dominate, and DeepMind has had a stellar year of success, but given its volume of output, OpenAI is really in a league of its own both in product impact, and in research that becomes quickly and broadly cited. The full top-100 list for 2022 is included below in this post.

Using data from the Zeta Alpha platform combined with careful human curation (more about methodology below), we've gathered the top cited papers in AI from 2022, 2021, and 2020, and analyzed authors' affiliations, and country. This allows us to rank these by R&D impact rather than pure publication volume.

What are some of these top papers we're talking about?

But before we dive into the numbers, let's get a sense of what papers we're talking about: the blockbusters from these past 3 years. You'll probably recognize a few of them!

1️⃣ AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models -> (From DeepMind, 1372 citations) Using AlphaFold to augment protein structure database coverage.

2️⃣ ColabFold: making protein folding accessible to all -> (From multiple institutions, 1162 citations) An open-source and efficient protein folding model.

3️⃣ Hierarchical Text-Conditional Image Generation with CLIP Latents -> (From OpenAI, 718 citations) DALL·E 2, complex prompted image generation that left most in awe.

4️⃣ A ConvNet for the 2020s -> (From Meta and UC Berkeley, 690 citations) A successful modernization of CNNs at a time of boom for Transformers in Computer Vision.

5️⃣ PaLM: Scaling Language Modeling with Pathways -> (From Google, 452 citations) Google's mammoth 540B Large Language Model, a new MLOps infrastructure, and how it performs.

1️⃣ Highly accurate protein structure prediction with AlphaFold -> (From DeepMind, 8965) AlphaFold, a breakthrough in protein structure prediction using Deep Learning. See also " Accurate prediction of protein structures and interactions using a three-track neural network " (from multiple academic institutions, 1659 citations), an open-source protein structure prediction algorithm.

2️⃣ Swin Transformer: Hierarchical Vision Transformer using Shifted Windows -> (From Microsoft, 4810 citations) A robust variant of Transformers for Vision.

3️⃣ Learning Transferable Visual Models From Natural Language Supervision -> (From OpenAI, 3204 citations) CLIP, image-text pairs at scale to learn joint image-text representations in a self-supervised fashion

4️⃣ On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? -> (From U. Washington, Black in AI, The Aether, 1266 citations) Famous position paper is very critical of the trend of ever-growing language models, highlighting their limitations and dangers.

5️⃣ Emerging Properties in Self-Supervised Vision Transformers -> (From Meta, 1219 citations) DINO, showing how self-supervision on images led to the emergence of some sort of proto-object segmentation in Transformers.

1️⃣ An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale -> (From Google, 11914 citations) The first work showing how a plain Transformer could do great in Computer Vision.

2️⃣ Language Models are Few-Shot Learners -> (From OpenAI, 8070 citations) GPT-3, This paper does not need further explanation at this stage.

3️⃣ YOLOv4: Optimal Speed and Accuracy of Object Detection -> (From Academia Sinica, Taiwan, 8014 citations) Robust and fast object detection sells like hotcakes.

4️⃣ Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer -> (From Google, 5906 citations) A rigorous study of transfer learning with Transformers, resulting in the famous T5.

5️⃣ Bootstrap your own latent: A new approach to self-supervised Learning -> (From DeepMind and Imperial College, 2873 citations) Showing that negatives are not even necessary for representation learning.

Read on below to see the full list of 100 papers for 2022, but let's first dive into the analyses for countries and institutions.

The most cited papers from the past 3 years

When we look at where these top-cited papers come from (Figure 1), we see that the United States continues to dominate and the difference among the major powers varies only slightly per year. Earlier reports that China may have overtaken the US in AI R&D seem to be highly exaggerated if we look at it from the perspective of citations. We also see an impact significantly above expectation from Singapore and Australia.

To properly assess the US dominance, let's look beyond paper count numbers. If we consider the accumulated citations by country instead, the difference looks even stronger. We have normalized by the total number of citations in a year, in order to be able to compare meaningfully across years.

Figure 2. Source: Zeta Alpha

The UK is clearly the strongest player outside of the US and China. However, the contribution of the UK is even more strongly dominated by DeepMind in 2022 (69% of the UK total), than in the previous years (60%). DeepMind has truly had a very productive 2022. Looking at the regions, North America is leading by a large margin while Asia is slightly above Europe.

Figure 3. Source: Zeta Alpha

Now let's look at how the leading organizations compare by number of papers in the top 100.

Figure 4. Source: Zeta Alpha

Google is consistently the strongest player followed by Meta, Microsoft, UC Berkeley, DeepMind and Stanford. While industry calls the shots in AI research these days, and single academic institutions don't produce as much impact, the tail for these institutions is much longer, so that when we aggregate by organization type, it evens out.

Figure 5. Source: Zeta Alpha

If we look into total research output, how many papers have organizations published in these past 3 years?

Figure 6. Source: Zeta Alpha

In total publication volume, Google is still in the lead, but differences are much less drastic compared to the citation top 100. You won't see OpenAI or DeepMind among the top 20 in the volume of publications. These institutions publish less but with higher impact. The following chart shows the rate at which organizations manage to convert their publications into top-100 papers.

Now we see that OpenAI is simply in a league of its own when it comes to turning publications into absolute blockbusters. While certainly, their marketing magic helps a lot to propel their popularity, it's undeniable that some of their recent research is of outstanding quality. With a lower paper volume but impressive conversion rate is also EleutherAI, the non-profit collective focusing on interpretability and alignment of large Language Models.

The top 100 most cited papers for 2022

And finally, here is our top-100 list itself, with titles, citation counts, and affiliations.

We have also added twitter mentions, which are sometimes seen as an early impact indicator, however the correlation so far seems to be weak. Further work is needed. Here you have the list for the year 2020 and for 2021 (as tsv files).

Methodology

To create the analysis above, we have first collected the most cited papers per year in the Zeta Alpha platform , and then manually checked the first publication date (usually an arXiv pre-print), so that we place papers in the right year. We supplemented this list by mining for highly cited AI papers on Semantic Scholar with its broader coverage and ability to sort by citation count. This mainly turns up additional papers from highly impactful closed-source publishers (e.g. Nature, Elsevier, Springer and other journals). We then take for each paper the number of citations on Google Scholar as the representative metric and sort the papers by this number to yield the top-100 for a year. For these papers we used GPT-3 to extract the authors, their affiliations, and their country and manually checked these results (if the country was not clearly visible from the publication, we take the country of the organization’s headquarters). A paper with authors from multiple affiliations counts once for each of the affiliations.

Updates 2023/03/07

- Update the 2022 list with the following papers:

- Emergent Abilities of Large Language Models (74 citations)

- Self-consistency improves chain of thought reasoning in language models (71 citations)

- Why do tree-based models still outperform deep learning on tabular data? (60 citations)

- DeiT III: Revenge of the ViT (44 citations)

- Fix missing EleutherAI as the 2nd best organization in terms of conversion rate

- Added links to the full 2020 top-cited paper list and 2021 top-cited paper lists

- Add counts by region plot

- Fix missing countries and organizations in the 2022 list

This concludes our analysis; what surprised you the most about these numbers? Try out our platform , follow us on Twitter @zetavector and let us know if you have any feedback or would like to receive a more detailed analysis for your domain or organization.

Recent Posts

Trends in AI — May 2024

Trends in AI — April 2024

Trends in AI — March 2024

Most Influential NATURE Papers (2022-02)

Nature is a British weekly scientific journal founded and based in London, England. As a multidisciplinary publication, Nature features peer-reviewed research from a variety of academic disciplines, mainly in science, technology, and the natural sciences. Paper Digest Team analyze all papers published on NATURE in the past years, and presents the 10 most influential papers for each year (based on when the paper became available online). This ranking list is automatically constructed based upon citations from both research papers and granted patents, and will be frequently updated to reflect the most recent changes. To find the most influential papers from other conferences/journals, visit Best Paper Digest page. Note: the most influential papers may or may not include the papers that won the best paper awards. (Last updated on: 2022-02-05)

If you do not want to miss any interesting academic paper, you are welcome to sign up our free daily paper digest service to get updates on new papers published in your area every day. To search for papers with highlights, related papers, patents, grants, experts and organizations, please visit our search console . You are also welcome to follow us on Twitter and Linkedin to get updated with new conference digests.

Paper Digest Team [email protected]

TABLE 1: Most Influential NATURE Papers (2022-02)

• Data Science
• Quantum Computing

• Miscellaneous

A Comprehensive Guide on RTMP Streaming

Blockchain booms, risks loom: the ai rescue mission in smart contract auditing, developing incident response plans for insider threats, weis wave: revolutionizing market analysis, top machine learning (ml) research papers released in 2022.

For every Machine Learning (ML) enthusiast, we bring you a curated list of the major breakthroughs in ML research in 2022.

Machine learning (ML) is gaining much traction in recent years owing to the disruption and development it brings in enhancing existing technologies. Every month, hundreds of ML papers from various organizations and universities get uploaded on the internet to share the latest breakthroughs in this domain. As the year ends, we bring you the Top 22 ML research papers of 2022 that created a huge impact in the industry. The following list does not reflect the ranking of the papers, and they have been selected on the basis of the recognitions and awards received at international conferences in machine learning.

• Bootstrapped Meta-Learning

Meta-learning is a promising field that investigates ways to enable machine learners or RL agents (which include hyperparameters) to learn how to learn in a quicker and more robust manner, and it is a crucial study area for enhancing the efficiency of AI agents.

This 2022 ML paper presents an algorithm that teaches the meta-learner how to overcome the meta-optimization challenge and myopic meta goals. The algorithm’s primary objective is meta-learning using gradients, which ensures improved performance. The research paper also examines the potential benefits due to bootstrapping. The authors highlight several interesting theoretical aspects of this algorithm, and the empirical results achieve new state-of-the-art (SOTA) on the ATARI ALE benchmark as well as increased efficiency in multitask learning.

• Competition-level code generation with AlphaCode

One of the exciting uses for deep learning and large language models is programming. The rising need for coders has sparked the race to build tools that can increase developer productivity and provide non-developers with tools to create software. However, these models still perform badly when put to the test on more challenging, unforeseen issues that need more than just converting instructions into code.

The popular ML paper of 2022 introduces AlphaCode, a code generation system that, in simulated assessments of programming contests on the Codeforces platform, averaged a rating in the top 54.3%. The paper describes the architecture, training, and testing of the deep-learning model.

• Restoring and attributing ancient texts using deep neural networks

The epigraphic evidence of the ancient Greek era — inscriptions created on durable materials such as stone and pottery —  had already been broken when it was discovered, rendering the inscribed writings incomprehensible. Machine learning can help in restoring, and identifying chronological and geographical origins of damaged inscriptions to help us better understand our past. 

This ML paper proposed a machine learning model built by DeepMind, Ithaca, for the textual restoration and geographical and chronological attribution of ancient Greek inscriptions. Ithaca was trained on a database of just under 80,000 inscriptions from the Packard Humanities Institute. It had a 62% accuracy rate compared to historians, who had a 25% accuracy rate on average. But when historians used Ithaca, they quickly achieved a 72% accuracy.

• Tensor Programs V: Tuning Large Neural Networks via Zero-Shot Hyperparameter Transfer

Large neural networks use more resources to train hyperparameters since each time, the network must estimate which hyperparameters to utilize. This groundbreaking ML paper of 2022 suggests a novel zero-shot hyperparameter tuning paradigm for more effectively tuning massive neural networks. The research, co-authored by Microsoft Research and OpenAI, describes a novel method called µTransfer that leverages µP to zero-shot transfer hyperparameters from small models and produces nearly perfect HPs on large models without explicitly tuning them.

This method has been found to reduce the amount of trial and error necessary in the costly process of training large neural networks. By drastically lowering the need to predict which training hyperparameters to use, this approach speeds up research on massive neural networks like GPT-3 and perhaps its successors in the future.

• PaLM: Scaling Language Modeling with Pathways 

Large neural networks trained for language synthesis and recognition have demonstrated outstanding results in various tasks in recent years. This trending 2022 ML paper introduced Pathways Language Model (PaLM), a 780 billion high-quality text token, and 540 billion parameter-dense decoder-only autoregressive transformer.

Although PaLM just uses a decoder and makes changes like SwiGLU Activation, Parallel Layers, Multi-Query Attention, RoPE Embeddings, Shared Input-Output Embeddings, and No Biases and Vocabulary, it is based on a typical transformer model architecture. The paper describes the company’s latest flagship surpassing several human baselines while achieving state-of-the-art in numerous zero, one, and few-shot NLP tasks.

• Robust Speech Recognition via Large-Scale Weak Supervision

Machine learning developers have found it challenging to build speech-processing algorithms that are trained to predict a vast volume of audio transcripts on the internet. This year, OpenAI released Whisper , a new state-of-the-art (SotA) model in speech-to-text that can transcribe any audio to text and translate it into several languages. It has received 680,000 hours of training on a vast amount of voice data gathered from the internet. According to OpenAI, this model is robust to accents, background noise, and technical terminology. Additionally, it allows transcription into English from 99 different languages and translation into English from those languages.

The OpenAI ML paper mentions the author ensured that about one-third of the audio data is non-English. This helped the team outperform other supervised state-of-the-art models by maintaining a diversified dataset.

• OPT: Open Pre-trained Transformer Language Models

Large language models have demonstrated extraordinary performance f on numerous tasks (e.g., zero and few-shot learning). However, these models are difficult to duplicate without considerable funding due to their high computing costs. Even while the public can occasionally interact with these models through paid APIs, complete research access is still only available from a select group of well-funded labs. This limited access has hindered researchers’ ability to comprehend how and why these language models work, which has stalled progress on initiatives to improve their robustness and reduce ethical drawbacks like bias and toxicity.

The popular 2022 ML paper introduces Open Pre-trained Transformers (OPT), a suite of decoder-only pre-trained transformers with 125 million to 175 billion parameters that the authors want to share freely and responsibly with interested academics. The biggest OPT model, OPT-175B (it is not included in the code repository but is accessible upon request), which is impressively proven to perform similarly to GPT-3 (which also has 175 billion parameters)  uses just 15% of GPT-3’s carbon footprint during development and training.

• A Path Towards Autonomous Machine Intelligence

Yann LeCun is a prominent and respectable researcher in the field of artificial intelligence and machine learning. In June, his much-anticipated paper “ A Path Towards Autonomous Machine Intelligence ” was published on OpenReview. LeCun offered a number of approaches and architectures in his paper that might be combined and used to create self-supervised autonomous machines. 

He presented a modular architecture for autonomous machine intelligence that combines various models to operate as distinct elements of a machine’s brain and mirror the animal brain. Due to the differentiability of all the models, they are all interconnected to power certain brain-like activities, such as identification and environmental response. It incorporates ideas like a configurable predictive world model, behavior-driven through intrinsic motivation, and hierarchical joint embedding architectures trained with self-supervised learning. 

• LaMDA: Language Models for Dialog Applications 

Despite tremendous advances in text generation, many of the chatbots available are still rather irritating and unhelpful. This 2022 ML paper from Google describes the LaMDA — short for “Language Model for Dialogue Applications” — system, which caused the uproar this summer when a former Google engineer, Blake Lemoine, alleged that it is sentient. LaMDA is a family of large language models for dialog applications built on Google’s Transformer architecture, which is known for its efficiency and speed in language tasks such as translation. The model’s ability to be adjusted using data that has been human-annotated and the capability of consulting external sources are its most intriguing features.

The model, which has 137 billion parameters, was pre-trained using 1.56 trillon words from publicly accessible conversation data and online publications. The model is also adjusted based on the three parameters of quality, safety, and groundedness.

• Privacy for Free: How does Dataset Condensation Help Privacy?

One of the primary proposals in the award-winning ML paper is to use dataset condensation methods to retain data efficiency during model training while also providing membership privacy. The authors argue that dataset condensation, which was initially created to increase training effectiveness, is a better alternative to data generators for producing private data since it offers privacy for free. 

Though existing data generators are used to produce differentially private data for model training to minimize unintended data leakage, they result in high training costs or subpar generalization performance for the sake of data privacy. This study was published by Sony AI and received the Outstanding Paper Award at ICML 2022. 

• TranAD: Deep Transformer Networks for Anomaly Detection in Multivariate Time Series Data

The use of a model that converts time series into anomaly scores at each time step is essential in any system for detecting time series anomalies. Recognizing and diagnosing anomalies in multivariate time series data is critical for modern industrial applications. Unfortunately, developing a system capable of promptly and reliably identifying abnormal observations is challenging. This is attributed to a shortage of anomaly labels, excessive data volatility, and the expectations of modern applications for ultra-low inference times. 

In this study , the authors present TranAD, a deep transformer network-based anomaly detection and diagnosis model that leverages attention-based sequence encoders to quickly execute inference while being aware of the more general temporal patterns in the data. TranAD employs adversarial training to achieve stability and focus score-based self-conditioning to enable robust multi-modal feature extraction. The paper mentions extensive empirical experiments on six publicly accessible datasets show that TranAD can perform better in detection and diagnosis than state-of-the-art baseline methods with data- and time-efficient training. 

• Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding 

In the last few years, generative models called “diffusion models” have been increasingly popular. This year saw these models capture the excitement of AI enthusiasts around the world. 

Going ahead of the current text to speech technology of recent times, this outstanding 2022 ML paper introduced the viral text-to-image diffusion model from Google, Imagen. This diffusion model achieves a new state-of-the-art FID score of 7.27 on the COCO dataset by combining the deep language understanding of transformer-based large language models with the photorealistic image-generating capabilities of diffusion models. A text-only frozen language model provides the text representation, and a diffusion model with two super-resolution upsampling stages, up to 1024×2014, produces the images. It employs several training approaches, including classifier-free guiding, to teach itself conditional and unconditional generation. Another important feature of Imagen is the use of dynamic thresholding, which stops the diffusion process from being saturated in specific areas of the picture, a behavior that reduces image quality, particularly when the weight placed on text conditional creation is large.

• No Language Left Behind: Scaling Human-Centered Machine Translation

This ML paper introduced the most popular Meta projects of the year 2022: NLLB-200. This paper talks about how Meta built and open-sourced this state-of-the-art AI model at FAIR, which is capable of translating 200 languages between each other. It covers every aspect of this technology: language analysis, moral issues, effect analysis, and benchmarking.

No matter what language a person speaks, accessibility via language ensures that everyone can benefit from the growth of technology. Meta claims that several languages that NLLB-200 translates, such as Kamba and Lao, are not currently supported by any translation systems in use. The tech behemoth also created a dataset called “FLORES-200” to evaluate the effectiveness of the NLLB-200 and show that accurate translations are offered. According to Meta, NLLB-200 offers an average of 44% higher-quality translations than its prior model.

• A Generalist Agent

AI pundits believe that multimodality will play a huge role in the future of Artificial General Intelligence (AGI). One of the most talked ML papers of 2022 by DeepMind introduces Gato – a generalist agent . This AGI agent is a multi-modal, multi-task, multi-embodiment network, which means that the same neural network (i.e. a single architecture with a single set of weights) can do all tasks while integrating inherently diverse types of inputs and outputs. 

DeepMind claims that the general agent can be improved with new data to perform even better on a wider range of tasks. They argue that having a general-purpose agent reduces the need for hand-crafting policy models for each region, enhances the volume and diversity of training data, and enables continuous advances in the data, computing, and model scales. A general-purpose agent can also be viewed as the first step toward artificial general intelligence, which is the ultimate goal of AGI. 

Gato demonstrates the versatility of transformer-based machine learning architectures by exhibiting their use in a variety of applications.  Unlike previous neural network systems tailored for playing games, stack blocks with a real robot arm, read words, and caption images, Gato is versatile enough to perform all of these tasks on its own, using only a single set of weights and a relatively simple architecture.

• The Forward-Forward Algorithm: Some Preliminary Investigations 

AI pioneer Geoffrey Hinton is known for writing paper on the first deep convolutional neural network and backpropagation. In his latest paper presented at NeurIPS 2022, Hinton proposed the “forward-forward algorithm,” a new learning algorithm for artificial neural networks based on our understanding of neural activations in the brain. This approach draws inspiration from Boltzmann machines (Hinton and Sejnowski, 1986) and noise contrast estimation (Gutmann and Hyvärinen, 2010). According to Hinton, forward-forward, which is still in its experimental stages, can substitute the forward and backward passes of backpropagation with two forward passes, one with positive data and the other with negative data that the network itself could generate. Further, the algorithm could simulate hardware more efficiently and provide a better explanation for the brain’s cortical learning process.

Without employing complicated regularizers, the algorithm obtained a 1.4 percent test error rate on the MNIST dataset in an empirical study, proving that it is just as effective as backpropagation.

The paper also suggests a novel “mortal computing” model that can enable the forward-forward algorithm and understand our brain’s energy-efficient processes.

• Focal Modulation Networks

In humans, the ciliary muscles alter the shape of the eye and hence the radius of the curvature lens to focus on near or distant objects. Changing the shape of the eye lens, changes the focal length of the lens. Mimicking this behavior of focal modulation in computer vision systems can be tricky.

This machine learning paper introduces FocalNet, an iterative information extraction technique that employs the premise of foveal attention to post-process Deep Neural Network (DNN) outputs by performing variable input/feature space sampling. Its attention-free design outperforms SoTA self-attention (SA) techniques in a wide range of visual benchmarks. According to the paper, focal modulation consists of three parts: According to the paper, focal modulation consists of three parts: 

a. hierarchical contextualization, implemented using a stack of depth-wise convolutional layers, to encode visual contexts from close-up to a great distance; 

b. gated aggregation to selectively gather contexts for each query token based on its content; and  

c. element-wise modulation or affine modification to inject the gathered context into the query.

• Learning inverse folding from millions of predicted structures

The field of structural biology is being fundamentally changed by cutting-edge technologies in machine learning, protein structure prediction, and innovative ultrafast structural aligners. Time and money are no longer obstacles to obtaining precise protein models and extensively annotating their functionalities. However, determining a protein sequence from its backbone atom coordinates remained a challenge for scientists. To date, machine learning methods to this challenge have been constrained by the amount of empirically determined protein structures available.

In this ICML Outstanding Paper (Runner Up) , authors explain tackling this problem by increasing training data by almost three orders of magnitude by using AlphaFold2 to predict structures for 12 million protein sequences. With the use of this additional data, a sequence-to-sequence transformer with invariant geometric input processing layers is able to recover native sequence on structurally held-out backbones in 51% of cases while recovering buried residues in 72% of cases. This is an improvement of over 10% over previous techniques. In addition to designing protein complexes, partly masked structures, binding interfaces, and numerous states, the concept generalises to a range of other more difficult tasks.

• MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge

Within the AI research community, using video games as a training medium for AI has gained popularity. These autonomous agents have had great success in Atari games, Starcraft, Dota, and Go. Although these developments have gained popularity in the field of artificial intelligence research, the agents do not generalize beyond a narrow range of activities, in contrast to humans, who continually learn from open-ended tasks.

This thought-provoking 2022 ML paper suggests MineDojo, a unique framework for embodied agent research based on the well-known game Minecraft. In addition to building an internet-scale information base with Minecraft videos, tutorials, wiki pages, and forum discussions, Minecraft provides a simulation suite with tens of thousands of open-ended activities. Using MineDojo data, the author proposes a unique agent learning methodology that employs massive pre-trained video-language models as a learnt reward function. Without requiring a dense shaping reward that has been explicitly created, MinoDojo autonomous agent can perform a wide range of open-ended tasks that are stated in free-form language.

• Is Out-of-Distribution Detection Learnable?

Machine learning (supervised ML) models are frequently trained using the closed-world assumption, which assumes that the distribution of the testing data will resemble that of the training data. This assumption doesn’t hold true when used in real-world activities, which causes a considerable decline in their performance. While this performance loss is acceptable for applications like product recommendations, developing an out-of-distribution (OOD) identification algorithm is crucial to preventing ML systems from making inaccurate predictions in situations where data distribution in real-world activities typically drifts over time (self-driving cars).

In this paper , authors explore the probably approximately correct (PAC) learning theory of OOD detection, which is proposed by researchers as an open problem, to study the applicability of OOD detection. They first focus on identifying a prerequisite for OOD detection’s learnability. Following that, they attempt to show a number of impossibility theorems regarding the learnability of OOD detection in a handful yet different scenarios.

• Gradient Descent: The Ultimate Optimizer 

Gradient descent is a popular optimization approach for training machine learning models and neural networks. The ultimate aim of any machine learning (neural network) method is to optimize parameters, but selecting the ideal step size for an optimizer is difficult since it entails lengthy and error-prone manual work. Many strategies exist for automated hyperparameter optimization; however, they often incorporate additional hyperparameters to govern the hyperparameter optimization process. In this study , MIT CSAIL and Meta researchers offer a unique approach that allows gradient descent optimizers like SGD and Adam to tweak their hyperparameters automatically.

They propose learning the hyperparameters by self-using gradient descent, as well as learning the hyper-hyperparameters via gradient descent, and so on indefinitely. This paper describes an efficient approach for allowing gradient descent optimizers to autonomously adjust their own hyperparameters, which may be layered recursively to many levels. As these gradient-based optimizer towers expand in size, they become substantially less sensitive to the selection of top-level hyperparameters, reducing the load on the user to search for optimal values.

• ProcTHOR: Large-Scale Embodied AI Using Procedural Generation 

Embodied AI is a developing study field that has been influenced by recent advancements in artificial intelligence, machine learning, and computer vision. This method of computer learning makes an effort to translate this connection to artificial systems. The paper proposes ProcTHOR, a framework for procedural generation of Embodied AI environments. ProcTHOR allows researchers to sample arbitrarily huge datasets of diverse, interactive, customisable, and performant virtual environments in order to train and assess embodied agents across navigation, interaction, and manipulation tasks. 

According to the authors, models trained on ProcTHOR using only RGB images and without any explicit mapping or human task supervision achieve cutting-edge results in 6 embodied AI benchmarks for navigation, rearrangement, and arm manipulation, including the ongoing Habitat2022, AI2-THOR Rearrangement2022, and RoboTHOR challenges. The paper received the Outstanding Paper award at NeurIPS 2022.

• A Commonsense Knowledge Enhanced Network with Retrospective Loss for Emotion Recognition in Spoken Dialog

Emotion Recognition in Spoken Dialog (ERSD) has recently attracted a lot of attention due to the growth of open conversational data. This is due to the fact that excellent speech recognition algorithms have emerged as a result of the integration of emotional states in intelligent spoken human-computer interactions. Additionally, it has been demonstrated that recognizing emotions makes it possible to track the development of human-computer interactions, allowing for dynamic change of conversational strategies and impacting the result (e.g., customer feedback). But the volume of the current ERSD datasets restricts the model’s development. 

This ML paper proposes a Commonsense Knowledge Enhanced Network (CKE-Net) with a retrospective loss to carry out dialog modeling, external knowledge integration, and historical state retrospect hierarchically. 

Subscribe to our newsletter

Subscribe and never miss out on such trending AI-related articles.

Join our WhatsApp Channel and Discord Server to be a part of an engaging community.

RELATED ARTICLES

Building high-quality datasets with llms, enhancing efficiency: the role of data storage in ai systems, from insight to impact: the power of data expanding your business, leave a reply cancel reply.

Save my name, email, and website in this browser for the next time I comment.

Most Popular

Analytics Drift strives to keep you updated with the latest technologies such as Artificial Intelligence, Data Science, Machine Learning, and Deep Learning. We are on a mission to build the largest data science community in the world by serving you with engaging content on our platform.

Contact us: [email protected]

Copyright © 2024 Analytics Drift Private Limited.

Suggestions or feedback?

MIT News | Massachusetts Institute of Technology

• Machine learning
• Social justice
• Black holes
• Classes and programs

Departments

• Aeronautics and Astronautics
• Brain and Cognitive Sciences
• Architecture
• Political Science
• Mechanical Engineering

Centers, Labs, & Programs

• Abdul Latif Jameel Poverty Action Lab (J-PAL)
• Picower Institute for Learning and Memory
• Lincoln Laboratory
• School of Architecture + Planning
• School of Engineering
• School of Humanities, Arts, and Social Sciences
• Sloan School of Management
• School of Science
• MIT Schwarzman College of Computing

MIT’s top research stories of 2022

Press contact :, media download, *terms of use:.

Images for download on the MIT News office website are made available to non-commercial entities, press and the general public under a Creative Commons Attribution Non-Commercial No Derivatives license . You may not alter the images provided, other than to crop them to size. A credit line must be used when reproducing images; if one is not provided below, credit the images to "MIT."

Previous image Next image

The dizzying pace of research and innovation at MIT can make it hard to keep up. To mark the end of the year, MIT News is looking back at 10 of the research stories that generated the most excitement in 2022.

We’ve also rounded up the year’s  top MIT community-related stories .

• Designing a heat engine with no moving parts . In April, engineers at MIT and the National Renewable Energy Laboratory (NREL) designed a heat engine that might someday enable a fully decarbonized power grid. In demonstrations, the engine was able to convert heat to electricity with over 40 percent efficiency — a performance better than that of traditional steam turbines.
• Creating a lightweight material stronger than steel . In February, MIT chemical engineers used a new polymerization process to form a material that that is stronger than steel and as light as plastic, and can be easily manufactured in large quantities. The material could be used as a coating for car parts or as a building material for bridges and other structures.
• Enabling portable desalination at the push of a button . MIT researchers developed a suitcase-sized device that can remove particles and salts to generate drinking water. Unlike other desalination units that rely on filters, this device uses electrical power to purify the water. It requires less power to operate than a cell phone charger and can be driven by a small solar panel. Just push start.
• Linking human genes to function . A team of researchers created the first map tying every gene expressed in human cells to its job in the cell. The map, which is available for other scientists to use, makes it easier to study a range of biological questions. The map was created using a CRISPR-based single-cell sequencing method known as Perturb-seq.
• Improving supercomputing with a new programming language . A team of researchers based mainly at MIT invented a faster and more reliable programming language for high-performance computing. The language, which was tested on a number of small programs, could one day help computers with a number of deep learning tasks like image processing.
• Lifting people out of extreme poverty . A study co-authored by an MIT economist showed that a one-time capital boost (in this case, a cow) helped poor people in rural Bangladesh improve their lives in the long run. The study suggests the very poor are in a poverty trap, in which an initial lack of resources prevents them from improving their circumstances, and implies that large asset transfers are an effective way to reduce global poverty.
• Helping robots fly . Inspired by fireflies, MIT researchers created tiny actuators that emit light to allow insect-scale robots to communicate. Weighing barely more than a paper clip, the robots are too small to make use of traditional means of sensing and communication. Instead, the actuators that control the robots’ wings light up in different colors and patterns, which could enable them to do things like share their location and call for help.
• Detecting a radio signal in a far-off galaxy . In July, astronomers at MIT and elsewhere were surprised to find a periodic fast radio burst (FRB) originating billions of light-years from Earth. It is the longest lasting FRB pattern detected to date and is made up of intensely strong radio waves that repeat every 0.2 seconds, similar to a heartbeat. Astronomers suspect the signal is coming from a neutron star.
• Proposal for a new, low-cost battery design . Researchers at MIT developed a battery made from abundant, inexpensive materials to complement the rise of lithium-ion batteries. The new battery uses aluminum and sulfur as its two electrode materials and a molten salt electrolyte in between. It could be ideal for powering single homes or small to medium sized businesses, producing a few tens of kilowatt-hours of storage capacity.
• Immigrants as job creators . A study co-authored by an MIT economist found that compared to native-born citizens, immigrants are about 80 percent more likely to found a firm. The study, which looked at registered businesses of all types across the country, suggests that immigrants act more as "job creators" than "job takers" and play outsized roles in high-growth entrepreneurship in the U.S.

Share this news article on:

Related topics.

• MIT Sloan School of Management
• School of Architecture and Planning
• School of Humanities Arts and Social Sciences

Related Articles

MIT community in 2022: A year in review

MIT’s top research stories of 2021

MIT community in 2021: A year in review

Previous item Next item

More MIT News

Understanding why autism symptoms sometimes improve amid fever

Read full story →

School of Engineering welcomes new faculty

Study explains why the brain can robustly recognize images, even without color

Turning up the heat on next-generation semiconductors

Sarah Millholland receives 2024 Vera Rubin Early Career Award

A community collaboration for progress

• More news on MIT News homepage →

Massachusetts Institute of Technology 77 Massachusetts Avenue, Cambridge, MA, USA

• Map (opens in new window)
• Events (opens in new window)
• People (opens in new window)
• Careers (opens in new window)
• Accessibility
• Social Media Hub
• MIT on Facebook
• MIT on YouTube
• MIT on Instagram

ROLAND GRAD, MD, MSc, AND MARK H. EBELL, MD, MS

Am Fam Physician. 2023;107(4):406-414

Author disclosure: Dr. Ebell is cofounder and editor-in-chief of Essential Evidence Plus; see Editor's Note . Dr. Grad has no relevant financial relationships.

This article summarizes the top 20 research studies of 2022 identified as POEMs (patient-oriented evidence that matters), excluding COVID-19. Statins for primary prevention of cardiovascular disease produce only a small absolute reduction in a person's likelihood of dying (0.6%), having a myocardial infarction (0.7%), or having a stroke (0.3%) over three to six years. Supplemental vitamin D does not reduce the risk of a fragility fracture, even in people with low baseline vitamin D levels or a previous fracture. Selective serotonin reuptake inhibitors are preferred medical therapy for panic disorder, and patients who discontinue antidepressants are more likely to relapse (number needed to harm = 6) compared with those who continue. Combination therapy using a selective serotonin reuptake inhibitor, serotonin-norepinephrine reuptake inhibitor, or tricyclic antidepressant with mirtazapine or trazodone is more effective than monotherapy for first-line treatment of acute severe depression and when monotherapy fails. Using hypnotic agents for insomnia in adults comes with a significant trade-off between effectiveness and tolerability. In patients with moderate to severe asthma, using a combination of albuterol and glucocorticoid inhalers as rescue therapy reduces exacerbations and need for systemic steroids. Observational research shows an increased risk of gastric cancer in patients taking proton pump inhibitors (number needed to harm = 1,191 over 10 years). The American College of Gastroenterology updated its guideline for gastroesophageal reflux disease, and a new guideline provides sound advice for the evaluation and management of irritable bowel syndrome. Adults older than 60 years with prediabetes are more likely to become normoglycemic than to develop diabetes mellitus or die. Treatment of prediabetes via intensive lifestyle intervention or metformin has no impact on long-term cardiovascular outcomes. Persons with painful diabetic peripheral neuropathy have similar degrees of improvement with monotherapy using amitriptyline, duloxetine, or pregabalin and greater improvement with combination therapy. When communicating with patients about disease risk, most patients prefer numbers over words because people overestimate word-based probabilities. In terms of drug therapy, the duration of an initial varenicline prescription should be 12 weeks. Many drugs can interact with cannabidiol. No important difference was found among ibuprofen, ketorolac, and diclofenac for treatment of acute nonradicular low back pain in adults.

For the past 24 years, a team of six clinicians has systematically reviewed more than 100 medical journals to find the research most likely to change and improve primary care practice. The team includes experts in family medicine, pharmacology, hospital medicine, and women's health. 1 , 2

The goal of this process is to identify POEMs (patient-oriented evidence that matters). A POEM must report at least one patient-oriented outcome, such as improvement in symptoms, morbidity, or mortality. It should also be free of important methodologic bias, making the results valid and trustworthy. Finally, if applied in practice, the results would change what some physicians do by adopting a new practice or discontinuing an old one shown to be ineffective or harmful. Of more than 20,000 research studies published in 2022 in the journals reviewed by the POEMs team, 253 met criteria for validity, relevance, and practice change. These POEMs are emailed daily to subscribers of Essential Evidence Plus (Wiley-Blackwell, Inc.).

The Canadian Medical Association purchases a POEMs subscription, and thousands of its members receive the daily POEM. These physicians can rate each one using a validated questionnaire. 3 This process is called the Information Assessment Method ( https://www.mcgill.ca/iam ). POEM ratings address the domains of clinical relevance, cognitive impact, use in practice, and expected health benefits. In 2022, each of the 253 daily POEMs were rated by an average of 1,087 physicians. New for 2022, readers of the daily POEM saw an “overuse alert” for POEMs that align with a Choosing Wisely recommendation. 4

In this article, the 12th installment of our annual series ( https://www.aafp.org/pubs/afp/content/top-poems.html ), we summarize the 20 most clinically relevant POEMs of 2022 as determined by Canadian Medical Association members. Although some of the most highly rated POEMs addressed COVID-19, rapid changes in management as the pandemic has evolved make many of them less relevant in 2023. Therefore, we briefly summarize these COVID-19 POEMs, as well as highly rated practice guidelines, separately. The full POEMs discussed in this review are available at https://www.aafp.org/pubs/afp/content/top-poems/2022.html .

Preventive Health Care

The first two POEMs relate to preventive medicine ( Table 1 ) . 5 , 6 Many patients take statins to reduce cardiovascular events. But how large is the benefit? A meta-analysis identified 19 studies that randomized 132,763 patients to statin therapy or placebo. 5 Studies were typically three to six years in duration, and statins reduced the risk of overall death by 0.8%, myocardial infarction by 1.3%, and stroke by 0.4%. When statins were used as primary prevention, the reductions were smaller: 0.6% fewer deaths, 0.7% fewer myocardial infarctions, and 0.3% fewer strokes, for an overall number needed to treat of 63 to prevent one event over three to six years. The 95% CI around all-cause mortality alone was 88 to 250 over several years of treatment as primary prevention. Not bad, but this is probably a higher number needed to treat than most of our patients believe.

Screening for low vitamin D levels and recommending replacement in those who are deficient have become popular, although the U.S. Preventive Services Task Force (USPSTF) found insufficient evidence to recommend screening in asymptomatic individuals. 7 A trial randomized 25,871 men 50 years or older and women 55 years or older to receive 2,000 IU of vitamin D or placebo daily for a median of five years. 6 There was no difference between groups in any type of fracture, even in patients with low baseline vitamin D levels (hazard ratio = 1.04; 95% CI, 0.80 to 1.36) or a previous fracture.

Behavioral Health

The first POEM in this group provides insight into what happens when a patient discontinues an antidepressant 8 ( Table 2 8 – 13 ) . The study included primary care patients with at least two episodes of depression who had been taking an antidepressant for at least two years. 8 Patients were randomized to continue the medication or taper it over two months to placebo. After one year, the likelihood of relapse was significantly higher in the discontinuation group (56% vs. 39%; hazard ratio = 2.1; 95% CI, 1.6 to 2.7). Symptoms in the discontinuation group were also more severe.

The next study looked at the best way to treat these patients with acute severe depression. It showed that combination therapy is better than monotherapy for first-line treatment and in patients who do not respond to initial monotherapy (standardized mean difference [SMD] = 0.31; 95% CI, 0.19 to 0.44). 9 An SMD of 0.31 is consistent with a small effect size. The combination of a monoamine reuptake inhibitor (selective serotonin reuptake inhibitor [SSRI], serotonin-norepinephrine reuptake inhibitor, or tricyclic antidepressant) with mirtazapine or trazodone is more effective than monotherapy for first-line treatment and for patients who do not respond to monotherapy.

Turning to anxiety disorders, a network meta-analysis of 87 randomized controlled trials with 12,800 participants evaluated 12 drug classes to treat panic disorder. 10 Only one trial was at low risk of bias, and most trials were no more than three months in duration. The meta-analysis found that the most effective drugs were benzodiazepines, tricyclic antidepressants, and SSRIs, with SSRIs having the least risk of adverse effects. Sertraline and escitalopram have the best balance of high remission rates and low risk of adverse events. But remember, other studies have found that cognitive behavior therapy and mindfulness-based therapies are likely to be as effective as medications. 14 , 15

Finally, a network meta-analysis identified 154 double-blind randomized controlled trials that compared a broad range of hypnotic agents with each other and placebo for the treatment of insomnia. 11 There were 20 individual drugs and drug classes, including benzodiazepines, melatonin, trazodone, and newer short-acting hypnotics such as eszopiclone (Lunesta) and lemborexant (Dayvigo). The drugs most likely to be effective were eszopiclone and lemborexant in the long-term and eszopiclone, lemborexant, benzodiazepines, seltorexant (undergoing clinical trials in the United States), trazodone, trimipramine, zopiclone (not available in the United States), and zolpidem in the short-term. However, all drugs had relatively high rates of adverse effects (although so did placebo). The authors concluded that eszopiclone and lemborexant provide the best balance of effectiveness and tolerability in the short- and long-term.

One POEM reported on a study of 3,132 children and adults with moderate to severe asthma who were taking long-term inhaled corticosteroids and had experienced a severe asthma exacerbation in the previous year 16 ( Table 3 16 , 17 ) . They were randomized to two puffs each of albuterol, 90 mcg, plus budesonide, 80 mcg; albuterol, 90 mcg, plus budesonide, or albuterol alone as a rescue inhaler. In the higher-dose budesonide group, there was about 1 fewer exacerbation per 7 patients treated for one year compared with the albuterol only group, and patients also took significantly fewer systemic steroids (84 mg vs. 130 mg per year).

The second POEM in this group adds to the same evidence base as the first but in a more diverse population. Asthma disproportionately affects Black and Latinx patients, yet these groups have been underrepresented in research. This study randomized 1,201 self-identified Black and Latinx adults with moderate to severe asthma to a short-acting beta agonist (SABA) inhaler alone as reliever therapy or to a SABA plus a puff of a beclomethasone, 80 mcg, inhaler for every puff of the SABA. 17 There was approximately 1 fewer asthma exacerbation for every 8 patients in the combined therapy group compared with SABA alone over one year. Asthma control scores also improved, but the change was not clinically significant. Minority representation in studies is critical to addressing health disparities. 18

Gastroenterology

The first of three POEMs on gastroenterology topics ( Table 4 19 – 22 ) summarized an observational study of the following question: Is there an association between gastric cancer and the use of proton pump inhibitors (PPIs)? Answer: Patients taking PPIs have a small increased risk of gastric cancer (number needed to harm = 1,191 over 10 years). 19 But, remember, association is not causation. That said, physicians initiating antacid therapy should begin with a histamine H 2 blocker. If physicians are prescribing a PPI, the lowest dose and duration possible should be used. Choosing Wisely Canada provides a toolkit for helping to deprescribe PPIs (available at https://choosingwiselycanada.org/toolkit/bye-bye-ppi ).

How should clinicians evaluate and manage patients with suspected gastroesophageal reflux disease? The American College of Gastroenterology guideline panel recognizes that some patients require long-term PPI therapy, and the benefits outweigh the theoretical risks in these patients. In patients who are treatment naive and have classic symptoms and no alarm symptoms, clinicians should attempt to discontinue PPIs after a successful eight-week trial. Therefore, initial prescriptions should generally be limited to a duration of eight weeks. 21

Informed by systematic reviews, the British Society of Gastroenterology published updated recommendations for the diagnosis of irritable bowel syndrome (defined as at least six months of abdominal pain or discomfort and altered bowel habits). 22 Colonoscopy is recommended only for patients with alarm signs and symptoms or those at risk of microscopic colitis. First-line therapy includes exercise and gradually increasing doses of soluble fiber (e.g., psyllium). Recommendations for other first- and second-line treatments are summarized in the POEM.

Prediabetes/Diabetes Mellitus

The natural history of prediabetes in adults older than 60 years is the topic of the first POEM in this category ( Table 5 23 – 26 ) . This English study included more than 2,000 adults without diabetes mellitus at baseline. 23 The average age of participants was 70.6 years, and 55% were women. Over eight years of follow-up, older people with prediabetes were more likely to become normoglycemic than to develop diabetes or die. Perhaps it is time to retire the term “prediabetes” when talking with older people.

The next POEM reports on the long-term cardiovascular outcomes for participants in the original Diabetes Prevention Program trial. 24 In this trial, participants were randomized to receive metformin, 850 mg twice daily; an intensive exercise program; or placebo, and were then followed for three years. At the conclusion of the study, all patients were invited to participate in a long-term open-label follow-up study, and 86% agreed. The bottom line: In patients with prediabetes, neither an intensive lifestyle intervention nor metformin had an impact on the long-term risk of cardiovascular outcomes.

In the nursing home setting, a cohort of 42 older adults with type 2 diabetes were evaluated for the frequency and persistence of hypoglycemia. All participants were taking medications capable of inducing hypoglycemia (e.g., sulfonylureas, repaglinide, insulin). The study showed that seven people (17%) spent more than 20% of an entire day in a hypoglycemic state. Severe hypoglycemia occurred in 19 people (45%). A lower A1C level was associated with more frequent events and a longer time spent in a hypoglycemic state. 25 We should allow A1C results to be higher in frail older patients.

The last POEM in this group summarized the findings of a crossover trial that compared three drugs for decreasing pain in adults with diabetic peripheral neuropathy. 26 Similar degrees of improvement were reported with monotherapy using amitriptyline, duloxetine (Cymbalta), or pregabalin (Lyrica). Greater improvement was reported with subsequent combination therapy, regardless of initial choice of medication. This POEM included a table summarizing the adverse effects associated with three combinations of these drugs. Dizziness was more common with duloxetine plus pregabalin, nausea with pregabalin plus amitriptyline, and dry mouth with amitriptyline plus pregabalin.

Miscellaneous

Five top POEMs do not fall easily into a single category ( Table 6 27 – 31 ) . With regard to health information, understanding should clearly precede decision-making. However, when we say “rare” or “high risk,” what are we implying, and what do these terms mean to patients? A systematic review of 33 studies found significant variability in the presentation of probabilities and their subsequent interpretation by patients. This POEM also showed that most patients prefer numbers rather than word-based estimates of risk. 27

The next POEM was a randomized trial evaluating whether extending varenicline (Chantix) treatment beyond 12 weeks is beneficial for increasing the likelihood of smoking cessation. At 52 weeks of follow-up, there was no difference in quit rates between those taking varenicline for 12 weeks vs. 24 weeks. 28 Also, combining varenicline with nicotine replacement therapy produced no additional benefit. Therefore, initial varenicline prescriptions should specify a duration of 12 weeks. Using varenicline for longer than 12 weeks may be appropriate for a select group of patients who want a longer treatment course. 32

Drug interactions between cannabidiol (CBD) and commonly used medications is the topic of the next POEM. This study showed that CBD may increase serum levels of SSRIs, tricyclic antidepressants, antipsychotics, beta blockers, and opioids. 29 CBD increases levels of lamotrigine, but other anti-convulsants decrease levels of CBD. Although this reminds us of the potential for drug-drug interactions with CBD, we also need to know more about the severity of these interactions for better clinical decision-making.

The next POEM addresses the treatment of acute nonradicular low back pain. A randomized controlled trial of 198 adults 18 to 65 years of age presenting to the emergency department found no difference among ibuprofen, ketorolac, and diclofenac for the outcome of clinical improvement at five days. 30 Thus, although some nonsteroidal anti-inflammatory drugs may have a reputation of being a better analgesic, there seems to be a class effect with these drugs.

Our final top POEM identified the best topical therapy for mild to moderate acne vulgaris. The authors of this network meta-analysis of randomized controlled trials began by asking patients what was most important to them in terms of the outcome of acne treatment. Thus, the primary outcomes of this study were effectiveness, the proportion of patients who reported at least moderate improvement, tolerability, and the proportion who withdrew from the study or stopped using the medication due to adverse events. The bottom line: For a condition such as acne that has many treatment choices, good options include adapalene/benzoyl peroxide (Epiduo), followed by clindamycin/benzoyl peroxide and then adapalene alone. 31

COVID-19 and Practice Guidelines

As noted earlier, studies of COVID-19 vaccines were among the most highly rated for relevance. The uniform message across these studies was that vaccines and boosters are highly effective, especially at preventing hospitalizations and in older patients. 33 – 37 They are also very safe. Although there is a small increase in the risk of myocarditis, especially in younger men (about 2 to 6 cases per 100,000 vaccinated men 18 to 39 years of age), the risk of myocarditis and other complications associated with COVID-19 infection and the lower risk of cardiac arrest or death in vaccinated people support the importance of vaccination. 38

Several guidelines were also highly rated. The USPSTF no longer recommends initiating aspirin therapy for primary prevention in patients 60 years or older based on newer studies showing no net benefit. The USPSTF continues to recommend statins for adults 40 to 75 years of age who have one or more cardiovascular risk factors and an estimated 10-year cardiovascular event risk of at least 10%, with shared decision-making for those with a 10-year event risk of 7.5% to 9.9%. 39

The National Institute for Health and Care Excellence updated guidelines for two important conditions. Although metformin remains the cornerstone of oral drug therapy for type 2 diabetes, increased use of sodium-glucose cotrans-porter-2 inhibitors is encouraged, especially for patients with chronic heart or renal disease. 40 The updated guideline includes a helpful decision tree (see https://www.bmj.com/content/377/bmj.o775/infographic ). The National Institute for Health and Care Excellence updated guidelines for depression highlight the importance of an ongoing patient-physician relationship and for management of less severe depression to begin with various nondrug treatment options, such as guided self-help. For more severe depression, treatment should begin with cognitive behavior therapy, medication, or both. 41

Editor's Note: This article was cowritten by Dr. Mark Ebell, deputy editor for evidence-based medicine for AFP and cofounder and editor-in-chief of Essential Evidence Plus, published by Wiley-Blackwell, Inc. Because of Dr. Ebell’s dual roles and ties to Essential Evidence Plus, the concept for this article was independently reviewed and approved by a group of AFP ’s medical editors. In addition, the article underwent peer review and editing by four of AFP ’s medical editors. Dr. Ebell was not involved in the editorial decision-making process.—Sumi Sexton, MD, Editor-in-Chief

The authors thank Wiley-Blackwell, Inc., for giving permission to excerpt the POEMs; Drs. Allen Shaughnessy, Henry Barry, David Slawson, Nita Kulkarni, and Linda Speer for selecting and writing the original POEMs; the academic family medicine fellows and faculty of the University of Missouri–Columbia for their work as peer reviewers; CMA Joule for supporting the POEMs CME program in Canada; Pierre Pluye, PhD, for codeveloping the Information Assessment Method; and Maria Vlasak for her assistance with copyediting the POEMs.

Shaughnessy AF, Slawson DC, Bennett JH. Becoming an information master: a guidebook to the medical information jungle. J Fam Pract. 1994;39(5):489-499.

Ebell MH, Barry HC, Slawson DC, et al. Finding POEMs in the medical literature. J Fam Pract. 1999;48(5):350-355.

Badran H, Pluye P, Grad R. When educational material is delivered: a mixed methods content validation study of the Information Assessment Method. JMIR Med Educ. 2017;3(1):e4.

CMA Joule medical librarians. Linking POEMs with overuse alerts from Choosing Wisely Canada. Canadian Medical Association. January 11, 2023. Accessed January 26, 2023. https://www.cma.ca/clinical-blog/linking-poems-overuse-alerts-choosing-wisely-canada

Byrne P, Demasi M, Jones M, et al. Evaluating the association between low-density lipoprotein cholesterol reduction and relative and absolute effects of statin treatment: a systematic review and meta-analysis [published correction appears in JAMA Intern Med . 2022; 182(5): 579]. JAMA Intern Med. 2022;182(5):474-481.

LeBoff MS, Chou SH, Ratliff KA, et al. Supplemental vitamin D and incident fractures in midlife and older adults. N Engl J Med. 2022;387(4):299-309.

U.S. Preventive Services Task Force. Final recommendation statement. Vitamin D deficiency in adults: screening. April 13, 2021. Accessed January 28, 2023. https://uspreventiveservicestaskforce.org/uspstf/recommendation/vitamin-d-deficiency-screening

Lewis G, Marston L, Duffy L, et al. Maintenance or discontinuation of antidepressants in primary care. N Engl J Med. 2021;385(14):1257-1267.

Henssler J, Alexander D, Schwarzer G, et al. Combining antidepressants vs antidepressant monotherapy for treatment of patients with acute depression: a systematic review and meta-analysis. JAMA Psychiatry. 2022;79(4):300-312.

Chawla N, Anothaisintawee T, Charoenrungrueangchai K, et al. Drug treatment for panic disorder with or without agoraphobia: systematic review and network meta-analysis of randomised controlled trials. BMJ. 2022;376:e066084.

De Crescenzo F, D'Alò GL, Ostinelli EG, et al. Comparative effects of pharmacological interventions for the acute and long-term management of insomnia disorder in adults: a systematic review and network meta-analysis. Lancet. 2022;400(10347):170-184.

Mysliwiec V, Martin JL, Ulmer CS, et al. The management of chronic insomnia disorder and obstructive sleep apnea: synopsis of the 2019 U.S. Department of Veterans Affairs and U.S. Department of Defense clinical practice guidelines [published correction appears in Ann Intern Med . 2021; 174(4): 584]. Ann Intern Med. 2020;172(5):325-336.

Qaseem A, Kansagara D, Forciea MA, et al. Management of chronic insomnia disorder in adults: a clinical practice guideline from the American College of Physicians. Ann Intern Med. 2016;165(2):125-133.

Hoge EA, Bui E, Mete M, et al. Mindfulness-based stress reduction vs escitalopram for the treatment of adults with anxiety disorders: a randomized clinical trial. JAMA Psychiatry. 2023;80(1):13-21.

Imai H, Tajika A, Chen P, et al. Psychological therapies versus pharmacological interventions for panic disorder with or without agoraphobia in adults. Cochrane Database Syst Rev. 2016(10):CD011170.

Papi A, Chipps BE, Beasley R, et al. Albuterol-budesonide fixed-dose combination rescue inhaler for asthma. N Engl J Med. 2022;386(22):2071-2083.

Israel E, Cardet JC, Carroll JK, et al. Reliever-triggered inhaled glucocorticoid in Black and Latinx adults with asthma. N Engl J Med. 2022;386(16):1505-1518.

Bryant-Stephens T. Breaking the skin color barriers for asthma medications. It's not black, brown, or white. N Engl J Med. 2022;386(16):1574-1575.

Abrahami D, McDonald EG, Schnitzer ME, et al. Proton pump inhibitors and risk of gastric cancer: population-based cohort study. Gut. 2022;71(1):16-24.

Seo SI, Park CH, You SC, et al. Association between proton pump inhibitor use and gastric cancer: a population-based cohort study using two different types of nationwide databases in Korea. Gut. 2021;70(11):2066-2075.

Katz PO, Dunbar KB, Schnoll-Sussman FH, et al. ACG clinical guideline for the diagnosis and management of gastroesophageal reflux disease. Am J Gastroenterol. 2022;117(1):27-56.

Vasant DH, Paine PA, Black CJ, et al. British Society of Gastroenterology guidelines on the management of irritable bowel syndrome. Gut. 2021;70(7):1214-1240.

Veronese N, Noale M, Sinclair A, et al. Risk of progression to diabetes and mortality in older people with prediabetes: The English longitudinal study on ageing. Age Ageing. 2022;51(2):afab222.

Goldberg RB, Orchard TJ, Crandall JP, et al. Effects of long-term metformin and lifestyle interventions on cardiovascular events in the Diabetes Prevention Program and its outcome study. Circulation. 2022;145(22):1632-1641.

Bouillet B, Tscherter P, Vaillard L, et al. Frequent and severe hypoglycaemia detected with continuous glucose monitoring in older institutionalised patients with diabetes. Age Ageing. 2021;50(6):2088-2093.

Tesfaye S, Sloan G, Petrie J, et al. Comparison of amitriptyline supplemented with pregabalin, pregabalin supplemented with amitriptyline, and duloxetine supplemented with pregabalin for the treatment of diabetic peripheral neuropathic pain (OPTION-DM): a multicentre, double-blind, randomised crossover trial [published correction appears in Lancet . 2022; 400(10355): 810]. Lancet. 2022;400(10353):680-690.

Andreadis K, Chan E, Park M, et al. Imprecision and preferences in interpretation of verbal probabilities in health: a systematic review. J Gen Intern Med. 2021;36(12):3820-3829.

Baker TB, Piper ME, Smith SS, et al. Effects of combined varenicline with nicotine patch and of extended treatment duration on smoking cessation: a randomized clinical trial. JAMA. 2021;326(15):1485-1493.

Balachandran P, Elsohly M, Hill KP. Cannabidiol interactions with medications, illicit substances, and alcohol: a comprehensive review. J Gen Intern Med. 2021;36(7):2074-2084.

Irizarry E, Restivo A, Salama M, et al. A randomized controlled trial of ibuprofen versus ketorolac versus diclofenac for acute, nonradicular low back pain. Acad Emerg Med. 2021;28(11):1228-1235.

Stuart B, Maund E, Wilcox C, et al. Topical preparations for the treatment of mild-to-moderate acne vulgaris: systematic review and network meta-analysis. Br J Dermatol. 2021;185(3):512-525.

Leone FT, Zhang Y, Evers-Casey S, et al. Initiating pharmacologic treatment in tobacco-dependent adults. An official American Thoracic Society clinical practice guideline. Am J Respir Crit Care Med. 2020;202(2):e5-e31.

Arbel R, Hammerman A, Sergienko R, et al. BNT162b2 vaccine booster and mortality due to Covid-19. N Engl J Med. 2021;385(26):2413-2420.

Magen O, Waxman JG, Makov-Assif M, et al. Fourth dose of BNT162b2 mRNA Covid-19 vaccine in a nationwide setting. N Engl J Med. 2022;386(17):1603-1614.

Talic S, Shah S, Wild H, et al. Effectiveness of public health measures in reducing the incidence of covid-19, SARS-CoV-2 transmission, and covid-19 mortality: systematic review and meta-analysis [published correction appears in BMJ . 2021; 375: n2997]. BMJ. 2021;375:e068302.

Altarawneh HN, Chemaitelly H, Ayoub HH, et al. Effects of previous infection and vaccination on symptomatic omicron infections. N Engl J Med. 2022;387(1):21-34.

Goldberg Y, Mandel M, Bar-On YM, et al. Protection and waning of natural and hybrid immunity to SARS-CoV-2. N Engl J Med. 2022;386(23):2201-2212.

Husby A, Hansen JV, Fosbøl E, et al. SARS-CoV-2 vaccination and myocarditis or myopericarditis: population based cohort study. BMJ. 2021;375:e068665.

Davidson KW, Barry MJ, Mangione CM, et al. Aspirin use to prevent cardiovascular disease: US Preventive Services Task Force Recommendation statement. JAMA. 2022;327(16):1577-1584.

Moran GM, Bakhai C, Song SH, et al. Type 2 diabetes: summary of updated NICE guidance. BMJ. 2022;377:o775.

National Institute for Health and Care Excellence. Depression in adults. June 29, 2022. Accessed January 28, 2023. https://www.nice.org.uk/guidance/ng222

Continue Reading

More in AFP

More in pubmed.

Copyright © 2023 by the American Academy of Family Physicians.

This content is owned by the AAFP. A person viewing it online may make one printout of the material and may use that printout only for his or her personal, non-commercial reference. This material may not otherwise be downloaded, copied, printed, stored, transmitted or reproduced in any medium, whether now known or later invented, except as authorized in writing by the AAFP.  See permissions  for copyright questions and/or permission requests.

Copyright © 2024 American Academy of Family Physicians. All Rights Reserved.

• Climate modelling
• Extreme weather
• Health and Security
• Temperature
• China energy
• Oil and gas
• Other technologies
• China Policy
• International policy
• Other national policy
• Rest of world policy
• UN climate talks
• Country profiles

Guest posts

• Infographics
• Media analysis
• State of the climate

Translations

• Daily Brief
• China Briefing
• Comments Policy
• Cookies Policy
• Global emissions
• Rest of world emissions
• UK emissions
• EU emissions
• Global South Climate Database
• Newsletters
• COP21 Paris
• COP22 Marrakech
• COP24 Katowice
• COP25 Madrid
• COP26 Glasgow
• COP27 Sharm el-Sheikh
• COP28 Dubai
• Privacy Policy
• Attribution
• Geoengineering
• Food and farming
• Nature policy
• Plants and forests
• Marine life
• Ocean acidification
• Ocean warming
• Sea level rise
• Human security
• Public health
• Public opinion
• Risk and adaptation
• Science communication
• Carbon budgets
• Climate sensitivity
• GHGs and aerosols
• Global temperature
• Negative emissions
• Rest of world temperature

Tipping points

• UK temperature
• Thank you for subscribing

Social Channels

Search archive.

Receive a Daily or Weekly summary of the most important articles direct to your inbox, just enter your email below. By entering your email address you agree for your data to be handled in accordance with our Privacy Policy .

Multiple Authors

Last year brought no shortage of news. Frontpages around the world were dominated by Russia’s invasion of Ukraine , the death of Queen Elizabeth II and a UK prime minister outlasted by a lettuce .

But in yet another hectic year for news coverage, climate change still made headlines – not least because of the thousands of peer-reviewed journal papers about climate and energy that are published every year. 

These studies were picked up around the world by online news outlets and shared on social media platforms, including Twitter, Facebook and LinkedIn. Tracking all these “mentions” was Altmetric , an organisation that scores academic papers according to the media attention they receive.

Using Altmetric data for 2022, Carbon Brief has compiled its now-traditional list of the 25 most talked-about climate or energy-related papers that were published the previous year.

From megafloods to megadroughts and insects to polar bears, last year saw a broad range of headline-grabbing research – as well as a new record-high Altmetric score for a paper in a Carbon Brief annual review.

The infographic above shows which papers made it into the top 10, while the chart at the end of the article shows which journals feature most frequently in the top 25.

Pandemic prominence

As in 2020 and 2021 , the most talked-about scientific research of the past year has been about Covid-19. 

All but five of the 50 highest scoring papers of 2022 relate to the coronavirus. Those five include two papers on monkeypox (since renamed “mpox”), one on the global burden of drug-resistant bacteria , one that identifies the Epstein-Barr virus as “the leading cause of multiple sclerosis”, and one finding that, in 2020, firearms became the main cause of deaths in children in the US.

Continuing the theme, the most talked-about climate and energy paper of 2022 also relates to the Covid-19 pandemic.

Just outside the overall top 50, in 56th place, is “ Climate change increases cross-species viral transmission risk ”. Published in Nature, the study warns that mammals forced to move to cooler climes amid rising global temperatures are “already” spreading their viruses further – with “undoubtable” impacts for human health.

The research uses modelling to map how climate change could shift the geographic ranges of 3,100 mammal species and the viruses they carry by 2070. It finds that climate change is increasingly driving new encounters between mammal species, raising the risk of novel disease spread. The world’s “biodiversity hotspots” and densely populated parts of Asia and Africa are most likely to be affected.

Speaking to Carbon Brief when the paper was published in April, co-lead author Dr Colin Carlson , a global change biologist at Georgetown University in Washington DC, explained:

“Species are going to show up in new combinations because of climate change and, when they do, that’s an opportunity for them to share viruses with each other.” 

The relevance of the work was “reinforced by the Covid-19 pandemic”, the paper says, “which began only weeks after the completion of this study”. The authors note that Covid-19 “probably originated in south-east Asian horseshoe bats and may have spread to humans through an as-yet-unknown bridge host”. They add:

“Although we caution against over-interpreting our results as explanatory of the current pandemic, our findings suggest that climate change could easily become the dominant anthropogenic force in viral cross-species transmission, which will undoubtedly have a downstream effect on human health and pandemic risk.”

The paper clocks in with an Altmetric score of 7,803 – the highest score for any climate paper featured in Carbon Brief’s annual reviews and only the second time the top-scoring paper has surpassed 7,000. 

(For Carbon Brief’s previous Altmetric articles, see the links for 2021 , 2020 , 2019 , 2018 , 2017 , 2016 and 2015 .)

The study was covered in 716 news stories by 558 outlets around the world, including the Guardian , New York Times , Agence France-Presse , Al Jazeera and New Delhi Times . It was also picked up in 44 blog posts.

The research was the most talked about on Twitter of any of the top 25 climate papers included here. The paper’s URL was included in more than 10,000 tweets from more than 9,000 accounts, which collectively have almost 39 million followers.

Considering the paper’s timing, it is “not surprising” that connecting climate change with pandemics “gathered a bit of attention”, Carlson tells Carbon Brief. 

However, he says, “there’s still a weird pattern where climate gets left out of the conversation [around pandemics] a lot compared to biodiversity loss or wildlife trade”. He adds:

“I think I’ve started to see that change finally, and if we played a part in that, that’s wonderful. Most of my work focuses on the health impacts of climate change, and I think time and time again, we underestimate how bad they’ll be…Climate change is an existential threat to human and animal health, and I hope our study and the conversation around it helped push that envelope a bit.”

The second highest-scoring climate paper of 2022 – clocking in at 70th in the overall 2022 rankings – is, “ Exceeding 1.5C global warming could trigger multiple climate tipping points ”. Published in the journal Science, the study has an Altmetric score of 6,573.

The study’s lead author, Dr David McKay , told Carbon Brief in September that tipping points have been a keen area of interest in the climate community since 2008, when the study, “ Tipping elements in the Earth’s climate system ”, first “broke the ice” on the subject.

His work provides the first comprehensive assessment of climate-related tipping points since the 2008 paper. It identifies 16 climate tipping elements – shown below – and finds a “significant likelihood” that multiple tipping points will be crossed if global temperatures exceed 1.5C above pre-industrial levels.

The study was mentioned in 667 news stories from 397 outlets – including the Guardian , New Scientist and BBC News . It received the highest number of mentions in blog posts and Wikipedia pages of the top 25 climate papers – at 55 and 34, respectively. The study also featured in more than 6,000 tweets.

The study was strategically published just days before researchers, economists and civil society representatives gathered in McKay’s home town of Exeter for a conference on the topic of climate tipping points. Carbon Brief attended the conference, and summarised the key talking points, ideas and proposals that emerged.

Arctic warming

In third place with an Altmetric score of 6,201 is the Communications Earth and Environment study, “ The Arctic has warmed nearly four times faster than the globe since 1979 ”.

It is well known that Arctic temperatures are rising much faster than the global average. Previous estimates suggest that the region is warming twice or even three times as quickly as the rest of the world. However, this new study finds that the Arctic has warmed nearly four times faster than the global average over the past four decades.

Dr Mika Rantanen – a researcher at the Finnish Meteorological Institute and lead author on the study – tells Carbon Brief that “Arctic warming and its consequences have become one of the biggest manifestations of climate change”. As such, he says he and his co-authors expected the paper to attract media attention.

True to his expectations, the article was picked up in 765 news stories from 525 outlets, including the Independent , New York Times , Scientific American and Washington Post . “I think that choosing an attractive title was one reason why the paper was picked up so widely,” Rantanen tells Carbon Brief.

In fourth place, with an Altmetric score of 6,147, is the Nature Climate Change paper, “ Greenland ice sheet climate disequilibrium and committed sea level rise ”. 

The study finds that the “imbalance” of the Greenland ice sheet caused by global warming means it is already committed to contributing “at least” 274mm to global sea levels in future, “regardless of 21st-century climate pathways”.

The research also warned that if Greenland melt was consistently as large as the high-melt year of 2012, this would commit 782mm to sea levels, which serves as “an ominous prognosis for Greenland’s trajectory through a 21st century of warming”.

The study was picked up by 867 news stories from 658 outlets. These include, for example, a Guardian headline warning that major sea level rise from Greenland was now “inevitable”, and an Associated Press article leading on the idea of “zombie ice”:

“That’s doomed ice that, while still attached to thicker areas of ice, is no longer getting replenished by parent glaciers now receiving less snow. Without replenishment, the doomed ice is melting from climate change and will inevitably raise seas.”

Study co-author Dr William Colgan , a glaciologist at the Geological Survey of Denmark and Greenland , told the outlet:

“It’s dead ice. It’s just going to melt and disappear from the ice sheet…This ice has been consigned to the ocean, regardless of what climate (emissions) scenario we take now.”

The paper was also picked up by 37 blog posts and 1,671 tweets.

Completing the top five is another pandemic-related paper, “ Over half of known human pathogenic diseases can be aggravated by climate change ”, also published in Nature Climate Change.

It finds that nearly 60% of pathogen-caused diseases that affect humans have been “at some point aggravated” by climate-related “hazards”, such as warming and drought. 

Specifically, the researchers identify how 375 infectious diseases around the world have been affected by a range of climatic hazards. They find that while 16% of infectious diseases have “at times” been reduced by climate hazards, there are more than 1,000 “unique pathways” by which pathogenic diseases were exacerbated by climate change. 

The authors warn that these pathways are “too numerous for comprehensive societal adaptations” and say that this work highlights the “urgent need to work at the source of the problem: reducing [greenhouse gas] emissions”.

With an overall Altmetric tally of 6,079, the paper would have scored high enough to secure top spot last year . It was covered by 803 news stories from 582 outlets, 31 posts from 28 blogs, and 2,571 tweets from 2,277 users.

Just missing out on a spot in the top five is a “perspective” paper in the Proceedings of the National Academy of Sciences, “ Climate endgame: Exploring catastrophic climate change scenarios ”. This less-than-cheery article on “catastrophic” outcomes of climate change was referenced in 556 news stories from 428 outlets. The BBC News coverage of the study says:

“Catastrophic climate change outcomes, including human extinction, are not being taken seriously enough by scientists, a new study says. The authors say that the consequences of more extreme warming – still on the cards if no action is taken – are ‘dangerously underexplored’. They argue that the world needs to start preparing for the possibility of what they term the ‘climate endgame’.”

With an Altmetric score of 4,807, spot number seven goes to the Advances in Atmospheric Sciences paper, “ Another record: Ocean warming continues through 2021 despite La Niña conditions ”.

The annual Lancet Countdown on Health and Climate Change report clocks in at number eight. The paper – under the heading “health at the mercy of fossil fuels” – was mentioned in 521 news stories from 386 outlets. In a striking summary, the report says:

“Worldwide, people are seeing their health increasingly affected by climate change amidst the compounding impacts of Covid-19 and the cost of living and energy crises; governments and companies continue to prioritise fossil fuels over a healthy future despite climate commitments; and rapid, holistic action is the only route to ensuring a just and healthy future.”

At number nine is the “brief communication” Nature Climate Change paper, “ Rapid intensification of the emerging south-western North American megadrought in 2020-2021 ”. The study finds that the American west’s megadrought is the driest in at least 1,200 years.

The research was picked up in 1,187 new stories – more than any of the other 25 papers included here – including American publications the New York Times , the Washington Post , NBC News and NPR . The paper also topped the tables for blog post mentions, with 76.

Rounding off the top 10 climate articles of 2022 is the Nature Climate Change study, “ Pronounced loss of Amazon rainforest resilience since the early 2000s ”, with an Altmetric score of 4,195. This article was mentioned in 563 news stories from 434 outlets, including Carbon Brief , BBC News , the Independent , the New York Times and New Scientist . 

The study finds that three-quarters of the Amazon rainforest has lost “resilience” since 2003 – making it more vulnerable to extreme events such as droughts. The work shows we are “approaching a tipping point”, the lead author of the study told journalists at a press conference.

Elsewhere in the top 25

The rest of the top 25 contains a varied mix of papers, including how agriculture and climate change are “ reshaping ” insect biodiversity worldwide (12th) and the discovery of a “ genetically distinct and functionally isolated ” population of polar bears from south-eastern Greenland (19th).

Just missing out on the top 10 is, “ Global carbon budget 2022 ”, in 11th place. Published in the journal Earth System Science Data, the paper details the annually estimated “global carbon budget” as produced by the Global Carbon Project .

The paper’s lead author Prof Pierre Friedlingstein , chair of mathematical modelling of climate systems at the University of Exeter , co-wrote a Carbon Brief article to unpack their findings.

The latest data shows that global CO2 emissions from fossil fuels and cement increased by 1.0% in 2022, hitting a new record high of 36.6bn tonnes of CO2. This rise was “primarily driven by a strong increase in oil emissions as global travel continues to recover from the Covid-19 pandemic”, the article says, adding that “coal and gas emissions grew more slowly, though both had record emissions in 2022”.

There are several energy-related papers in the top 25. This includes a paper in 15th place from the Proceedings of the National Academy of Sciences , which finds that “corn-based ethanol in the US has failed to meet the policy’s own greenhouse gas emissions targets and negatively affected water quality, the area of land used for conservation, and other ecosystem processes”. Another is a Joule paper in 20th, which finds that “compared to continuing with a fossil fuel-based system, a rapid green energy transition is likely to result in trillions of net savings”.

The topic of public health features again in the top 25. For example, in 14th place is a Communications Earth & Environment study warning that, under 2C of warming, the exposure to dangerous heat stress “will likely increase by 50-100% across much of the tropics and increase by a factor of 3-10 in many regions throughout the mid-latitudes”.

Expert analysis direct to your inbox.

Get a round-up of all the important articles and papers selected by Carbon Brief by email. Find out more about our newsletters here .

Research into extreme weather also appears in the top 25. This includes a Nature Communications study in 18th place on the 2020 North Atlantic hurricane season – the most active on record – which finds that “human-induced climate change increased the extreme three-hourly storm rainfall rates and extreme three-day accumulated rainfall amounts…for observed storms that are at least tropical storm strength”. A Science Advances paper in 23rd place warns that climate change has already doubled the likelihood of an event capable of producing a catastrophic “megaflood” in California.

Finally, rounding out the top 25 is a Nature Geoscience study showing that the Thwaites glacier in west Antarctica has seen “sustained pulses of rapid retreat” in the past two centuries and similar pulses “are likely to occur in the near future”.

All the final scores for the top 25 climate papers of 2022 can be found in this spreadsheet .

Top journals

Across the top 25 papers in Carbon Brief’s leaderboard this year, Nature Climate Change features most frequently with four papers. Nature Climate Change also took first place in 2021 (jointly with Nature) and 2016 (jointly with Science).  

In joint-second place is Nature and Science with three papers each. Nature is perennially high-placed in this analysis, taking first – or joint first – spot in Carbon Brief’s top 25 in 2021 , 2020 , 2019 , 2018 , 2017 and 2015 .

For the rest of the top 25, there are four journals that appear twice and seven that appear once.

Diversity of the top 25

The top 25 climate papers of 2022 cover a huge range of topics and scope. 

The “Lancet Countdown on health and climate change” and “Global Carbon Budget” reports are epic annual publications, which review vast swathes of literature and have around 100 authors each. Meanwhile, three of the climate papers in the list have only two authors, and focus on specific new frontiers of climate research.

In total, the top 25 climate papers of 2022 have more than 400 authors. However, despite the variety in the climate research the papers present, analysis of their authors reveals a distinct lack of diversity.

Carbon Brief recorded the gender and country of affiliation for each of these authors. (The  methodology used was developed by Carbon Brief for analysis presented in a special 2021 series on climate justice .) The analysis reveals that the authors of the climate papers most featured in the media in 2022 are predominantly men from the global north.

The chart below shows the institutional affiliations of all authors in this analysis, broken down by continent – Europe, North America , Oceania, Asia, South America and Africa.

The analysis shows that nine out of every 10 authors are affiliated with institutions from the global north – defined as North America, Europe and Oceania. Meanwhile, there are only two authors from South America, based in Peru, and a single author from Africa, based in South Africa.

Further data analysis shows that there are also inequalities within continents. The map below shows the percentage of authors from each country in the analysis, where dark blue indicates a higher percentage. Countries that are not represented by any authors in the analysis are shown in white.

The top-ranking countries on this map are the US and the UK, which together account for more than half of all authors in this analysis (32% and 19%, respectively). Almost half of all researchers from the global south are from China – which accounts for around 6% of all researchers in the analysis.

None of the authors of the top 25 climate papers of 2022 is from south Asia, central America or the Caribbean.

Meanwhile, only 27% of authors from the top 25 climate papers of 2022 are women. Similarly, only six of the 25 papers have a female lead author.

The plot below shows the number of male (purple) and female (orange) authors in this analysis from each continent.

The full spreadsheet showing the results of this data analysis can be found here. For more on the biases in climate publishing, see Carbon Brief’s article on the lack of diversity in climate-science research .

嘉宾来稿：满足中国增长的用电需求 光伏加储能“比新建煤电更实惠”

DeBriefed 24 May 2024: ‘Surprise’ UK election; Oceans court ruling; China and Russia’s fossil-fuel pact

Guest post: Heat pumps gained European market share in 2023 despite falling sales

DeBriefed 17 May 2024: Biden’s clean-energy tariff blitz; Modi’s coal plans examined; Deadly heat in Mexico

Update: This article was updated on 09/01/2023 to add in the diversity analysis.

• Analysis: The climate papers most featured in the media in 2022

• Conferences
• Last updated October 25, 2022
• In AI Origins & Evolution

Top 10 Machine Learning Papers of 2022

• Published on October 25, 2022
• by Tasmia Ansari

The relevance of any field depends on the ongoing research and studies around it. This especially holds for advancing fields like machine learning. 

To bring you up to speed on the critical ideas driving machine learning in 2022, we handpicked the top 10 research papers for all AI/ML enthusiasts out there!

Let’s dive in!

• Artificial Replay: A Meta-Algorithm for Harnessing Historical Data in Bandits

Author(s) – Sean R. Sinclair et al .

Ways to incorporate historical data are still unclear: initialising reward estimates with historical samples can suffer from bogus and imbalanced data coverage, leading to computational and storage issues—particularly in continuous action spaces. The paper addresses the obstacles by proposing ‘Artificial Replay’, an algorithm to incorporate historical data into any arbitrary base bandit algorithm. 

Read the full paper here . 

• Bootstrapped Meta-Learning 

Author(s) – Sebastian Flennerhag et al.

The paper proposes an algorithm in which the meta-learner teaches itself to overcome the meta-optimisation challenge. The algorithm focuses on meta-learning with gradients, which guarantees performance improvements. Furthermore, the paper also looks at how bootstrapping opens up possibilities. 

Read the full paper here .

• LaMDA: Language Models for Dialog Applications

Author(s) – Romal Thoppilan et al.

The research describes the LaMDA system which caused chaos in AI this summer when a former Google engineer claimed that it had shown signs of sentience. LaMDA is a family of large language models for dialogue applications based on Transformer architecture. The interesting feature of the model is its fine-tuning with human-annotated data and the possibility of consulting external sources. This is a very interesting model family, which we might encounter in many applications we use daily. 

• Competition-Level Code Generation with AlphaCode

Author(s) – Yujia Li et al.

Systems can help programmers become more productive. The following research addresses the problems with incorporating innovations in AI into these systems. AlphaCode is a system that creates solutions for problems that require deeper reasoning. 

• Privacy for Free: How does Dataset Condensation Help Privacy?

Author(s) – Tian Dong et al.

The paper focuses on Privacy Preserving Machine Learning, specifically deducting the leakage of sensitive data in machine learning. It puts forth one of the first propositions of using dataset condensation techniques to preserve the data efficiency during model training and furnish membership privacy.

• Why do tree-based models still outperform deep learning on tabular data?

Author(s) – Léo Grinsztajn, Edouard Oyallon and Gaël Varoquaux

The research answers why deep learning models still find it hard to compete on tabular data compared to tree-based models. It is shown that MLP-like architectures are more sensitive to uninformative features in data compared to their tree-based counterparts. 

• Multi-Objective Bayesian Optimisation over High-Dimensional Search Spaces 

Author(s) – Samuel Daulton et al.

The paper proposes ‘MORBO’, a scalable method for multiple-objective BO as it performs better than that of high-dimensional search spaces. MORBO significantly improves the sample efficiency and, where existing BO algorithms fail, MORBO provides improved sample efficiencies over the current approach. 

• A Path Towards Autonomous Machine Intelligence Version 0.9.2

Author(s) – Yann LeCun

The research offers a vision about how to progress towards general AI. The study combines several concepts: a configurable predictive world model, behaviour driven through intrinsic motivation, and hierarchical joint embedding architectures trained with self-supervised

learning. 

• TranAD: Deep Transformer Networks for Anomaly Detection in Multivariate Time Series Data

Author(s) –   Shreshth Tuli, Giuliano Casale and Nicholas R. Jennings

This is a specialised paper applying transformer architecture to the problem of unsupervised anomaly detection in multivariate time series. Many architectures which were successful in other fields are, at some point, also being applied to time series. The research shows improved performance on some known data sets. 

• Differentially Private Bias-Term only Fine-tuning of Foundation Models

Author(s) – Zhiqi Bu et al. 

In the paper, researchers study the problem of differentially private (DP) fine-tuning of large pre-trained models—a recent privacy-preserving approach suitable for solving downstream tasks with sensitive data. Existing work has demonstrated that high accuracy is possible under strong privacy constraints yet requires significant computational overhead or modifications to the network architecture.

Read the full paper here . 

Access all our open Survey & Awards Nomination forms in one place

Tasmia Ansari

A ‘Jio Moment’ in AI’ has been long overdue.

Top Editorial Picks

Vishal Bhola Appointed President of Nothing India Mohit Pandey

Zoho Invests in Drone Startup Yali Aerospace to Solve Emergency Medical Deliveries  Vandana Nair

TCS & IIT-Bombay to Build India’s First Quantum Diamond Microchip Imager Pritam Bordoloi

Apple Introduces Denoising LM for Correcting Errors in ASR Systems Mohit Pandey

Subscribe to The Belamy: Our Weekly Newsletter

Biggest ai stories, delivered to your inbox every week..

"> "> Flagship Events

Join the forefront of data innovation at the Data Engineering Summit 2024, where industry leaders redefine technology’s future.

© Analytics India Magazine Pvt Ltd & AIM Media House LLC 2024

• Terms of use
• Privacy Policy

Subscribe to our Youtube channel and see how AI ecosystem works.

Main Navigation

• Contact NeurIPS
• Code of Ethics
• Code of Conduct
• Create Profile
• Journal To Conference Track
• Diversity & Inclusion
• Proceedings
• Future Meetings
• Exhibitor Information
• Privacy Policy

NeurIPS 2024, the Thirty-eighth Annual Conference on Neural Information Processing Systems, will be held at the Vancouver Convention Center

Monday Dec 9 through Sunday Dec 15. Monday is an industry expo.

Registration

Pricing » Registration 2024 Registration Cancellation Policy » . Certificate of Attendance

Our Hotel Reservation page is currently under construction and will be released shortly. NeurIPS has contracted Hotel guest rooms for the Conference at group pricing, requiring reservations only through this page. Please do not make room reservations through any other channel, as it only impedes us from putting on the best Conference for you. We thank you for your assistance in helping us protect the NeurIPS conference.

Announcements

• The call for High School Projects has been released
• The Call For Papers has been released
• See the Visa Information page for changes to the visa process for 2024.

Latest NeurIPS Blog Entries [ All Entries ]

Important dates.

If you have questions about supporting the conference, please contact us .

View NeurIPS 2024 exhibitors » Become an 2024 Exhibitor Exhibitor Info »

Organizing Committee

General chair, program chair, workshop chair, workshop chair assistant, tutorial chair, competition chair, data and benchmark chair, diversity, inclusion and accessibility chair, affinity chair, ethics review chair, communication chair, social chair, journal chair, creative ai chair, workflow manager, logistics and it, mission statement.

The Neural Information Processing Systems Foundation is a non-profit corporation whose purpose is to foster the exchange of research advances in Artificial Intelligence and Machine Learning, principally by hosting an annual interdisciplinary academic conference with the highest ethical standards for a diverse and inclusive community.

About the Conference

The conference was founded in 1987 and is now a multi-track interdisciplinary annual meeting that includes invited talks, demonstrations, symposia, and oral and poster presentations of refereed papers. Along with the conference is a professional exposition focusing on machine learning in practice, a series of tutorials, and topical workshops that provide a less formal setting for the exchange of ideas.

More about the Neural Information Processing Systems foundation »

• All Solutions
• Audience measurement
• Media planning
• Marketing optimization
• Content metadata
• Nielsen One
• All Insights
• Case Studies
• Perspectives
• Data Center
• The Gauge TM – U.S.
• Top 10 – U.S.
• Top Trends – Denmark
• Top Trends – Germany
• Women’s World Cup
• Men’s World Cup
• Big Data + Panel
• News Center

Client Login

2024 Annual Marketing Report

Discover how global marketers are allocating budgets, maximizing ROI and what these trends mean for your own impact…

Are you investing in performance marketing for the right reasons?

A look at how ctv reach and viewership trends shift across generations.

Influencer marketing: The obvious approach

Working with Nielsen’s Brand Impact solution has been a valuable partnership for Obviously,” says Heather at…

‘Data driven’ is no longer enough for your ROI strategy

ROI strategies hinge on capturing the right data at every stage of the customer journey. Just because data is easy to…

Outcomes-minded metrics: The marketing KPIs your CFO cares about

These are five outcomes-focused KPIs that help marketers show impact and unlock budget.

Reaching Asian American Audiences 2024

Understanding the media preferences of Asian American, Native Hawaiian and Pacific Islanders is critical to resonating in…

Featured reports

Metadata matters: Powering future FAST channel success

This guide will help FAST channels prepare for the future, when search and discovery features within individual services…

Explore all insights

Find the right solution for your business

In an ever-changing world, we’re here to help you stay ahead of what’s to come with the tools to measure, connect with, and engage your audiences.

How can we help?