ClaimsPeople

Yann LeCun

Yann LeCun is a computer scientist who shared the 2018 ACM Turing Award, is a Jacob T. Schwartz Professor at NYU's Courant Institute, was Meta's Chief AI Scientist and founding director of FAIR, and is Executive Chairman of AMI Labs.

1,175 claims collected, 197 researched, 84 resolved. Report of 4 November 2026, method 0.1.0+c876030c.

Scores by kind of claim

Each kind of claim is scored on its own, out of 100: the share that held up, with the range it likely falls in. Fewer resolved claims give a wider range.

  • Factual claims

    Statements about the present or the past, checked against the record: did they hold up?

    78 72 83

    High confidence, 79 factual claims resolved. From 16 documents, 2006 to 2026; 19 documents sampled from 33 candidates found.

    Not scored: 43 unresolvable.

  • Predictions

    Statements about events outside the person's control, checked once they fell due: did they come true?

    64 43 83

    Low confidence, 5 predictions resolved. From 4 documents, 2010 to 2024; 19 documents sampled from 33 candidates found.

    Not scored: 66 unresolvable. 3 more fall due later.

    Does not yet allow for how likely each prediction was when made.

  • Promises

    Commitments to the person's or organisation's own future action, with something to deliver: was it delivered?

    Insufficient record: 0 promises resolved, 5 needed for a score.

    Not scored: 1 unresolvable.

All types combined: 77 (likely 71 to 82, high confidence), from 84 claims: 79 factual claims, 5 predictions. It mixes kinds of claim in different proportions for different people, so compare people kind by kind.

Yann LeCun's claims

197 claims, sorted by newest said. Choose a claim to read it in full.
VerdictKindSaidClaimTopic
Mostly true Factual 2021 In reinforcement learning the machine receives only a single scalar feedback, and only occasionally. Factual · 2021 AI capabilities
Unresolvable Prediction 2021 All of intelligence could possibly be solved by filling in blanks in language and similar self-supervised prediction. Prediction · 2021 AI timelines and forecasting
Unresolvable Factual 2021 AI is not anywhere close to reproducing the learning done by a cat brain of about 800 million neurons. Factual · 2021 AI capabilities
Unresolvable Factual 2021 Most of the complicated components of intelligence already exist in cats, dogs and non-human primates. Factual · 2021 Science and research methods
Unresolvable Prediction 2021 Active learning and reinforcement learning will become obsolete, useless or easy once self-supervised representation learning or predictive world models are figured out. Prediction · 2021 AI capabilities
Mostly true Factual 2021 The amount of image training data available for self-supervised learning is essentially unlimited, with billions of new images per day. Factual · 2021 AI compute, hardware and data
Unresolvable Factual 30 Sep 2020 An implementation of Perlin noise in PyTorch has been made by the X user honualx. Factual · 30 Sep 2020 Technology
Unresolvable Factual 30 Sep 2020 Ken Perlin is a colleague of Yann LeCun at NYU. Factual · 30 Sep 2020 Their own work or organisation
Unresolvable Factual 29 Sep 2020 The final lecture of the previous semester's Deep Learning course has been posted online. Factual · 29 Sep 2020 Field building and careers
Unresolvable Factual 28 Sep 2020 Konbini published a short interview with Yann LeCun on the history of deep learning and his career. Factual · 28 Sep 2020 Their own work or organisation
Resolved true Factual 31 Aug 2020 Facebook AI released Opacus, a new PyTorch library for machine learning with differential privacy. Factual · 31 Aug 2020 Technology
Unresolvable Factual 19 Aug 2020 Africa has been developing quickly over the last decades. Factual · 19 Aug 2020 Economics and markets
Resolved true Factual 18 Aug 2020 The FastMRI project is a two-year-old collaboration between NYU Langone Health and Facebook. Factual · 18 Aug 2020 Biosecurity
Unresolvable Factual 14 Aug 2020 The Wall Street Journal published a piece on self-supervised learning involving Kyunghyun Cho and Yann LeCun. Factual · 14 Aug 2020 Their own work or organisation
Unresolvable Factual 9 Aug 2020 The double descent test-error peak does not appear with multilayer neural networks because they are implicitly regularized. Factual · 9 Aug 2020 Science and research methods
Mostly true Factual 28 Jun 2019 Jumping spiders can count to three, with categories for one prey, two preys and many preys. Factual · 28 Jun 2019 Other
Mostly true Factual 25 Jun 2019 Nick Clegg, Facebook's VP of global affairs and communications, wrote an op-ed about the future of the Internet. Factual · 25 Jun 2019 AI policy and governance
Resolved true Factual 21 Jun 2019 A new XLM model outperforms BERT on all GLUE tasks when trained on the same data. Factual · 21 Jun 2019 AI capabilities
Unresolvable Factual 18 Jun 2019 Calibra is a Facebook subsidiary and the first member of Libra. Factual · 18 Jun 2019 Economics and markets
Unresolvable Factual 18 Jun 2019 French fries originated in France. Factual · 18 Jun 2019 Other
Unresolvable Factual 11 Jun 2019 The figure given for PhD in Data Science is 47%. Factual · 11 Jun 2019 Field building and careers
Resolved true Factual 10 Jun 2019 PyTorch Hub, a repository offering pretrained models such as ResNet, BERT and GPT, was announced on 2019-06-10. Factual · 10 Jun 2019 Technology
Unresolvable Factual 9 Jun 2019 In US facilities holding immigrant children, in violation of the law, no services are provided beyond food, water and very basic shelter. Factual · 9 Jun 2019 Philosophy, ethics and society
Unresolvable Factual 9 Jun 2019 The US operates concentration camps for immigrant children. Factual · 9 Jun 2019 Philosophy, ethics and society
Mostly true Factual 3 Jun 2019 The Beijing Academy of Artificial Intelligence's published AI ethics guidelines mention the protection of individual privacy. Factual · 3 Jun 2019 AI policy and governance
Resolved true Factual 2019 In computer vision, the trend is to apply ConvNets to an entire image to detect, segment and recognize objects of any size at any location. Factual · 2019 AI capabilities
Unresolvable Factual 2019 Current model-free RL methods would likely need millions of hours of practice, with numerous accidents, for an autonomous car to learn to drive. Factual · 2019 AI capabilities
Unresolvable Factual 2019 Training a video predictor with a least-square criterion yields blurry frames that average all possible futures. Factual · 2019 AI capabilities
Unresolvable Prediction 2019 The demands on DL-specific hardware will undoubtedly increase. Prediction · 2019 AI compute, hardware and data
Unresolvable Prediction 2019 Chances are that the bulk of computation in future deep learning systems will still consist primarily of convolutions. Prediction · 2019 AI compute, hardware and data
Unresolvable Prediction 2019 Programmable resistor technology may become sufficiently compact, reliable, durable and configurable for deep learning applications. Prediction · 2019 AI compute, hardware and data
Unresolvable Prediction 2019 Augmented/virtual reality and telepresence applications will require extremely low-power ASICs for DL inference. Prediction · 2019 AI compute, hardware and data
Mostly true Factual 2019 The ConvNet architecture was loosely inspired by the ventral pathway in the visual cortex. Factual · 2019 AI capabilities
Mostly true Factual 2019 ReLU nonlinearities and dropout regularization enabled purely supervised backprop training of very deep networks with a dozen layers. Factual · 2019 AI capabilities
Unresolvable Factual 2019 Dataflow stream processors have become a popular design for ConvNet accelerators, such as the Eyeriss project. Factual · 2019 AI compute, hardware and data
Unresolvable Prediction 2019 The ultimate solution to DL power constraints may be exploiting sparse activations, perhaps with event-based computation. Prediction · 2019 AI compute, hardware and data
Unresolvable Prediction 2019 What the AI hardware and software communities produce over the next few years will shape AI research for the coming decades. Prediction · 2019 AI compute, hardware and data
Unresolvable Prediction 2019 Self-supervised latent-variable predictive models will be the centerpiece of intelligent systems based on model-predictive control and model-based reinforcement learning, e.g. for robotic manipulation and autonomous driving. Prediction · 2019 AI capabilities
Mostly false Factual 2019 Users upload 2 to 3 billion photos on Facebook every day. Factual · 2019 AI compute, hardware and data
Mostly true Prediction 2019 Usage of graph neural networks will increase for a wide variety of applications likely to violate the assumptions of current deep learning hardware. Prediction · 2019 AI compute, hardware and data
Unresolvable Prediction 2019 New high-performance hardware will be required to enable progress towards deep learning systems trained to acquire vast amounts of background knowledge and a form of common sense. Prediction · 2019 AI compute, hardware and data
Mostly true Prediction 2019 As DL is used for high-level cognitive tasks, memory modules will become commonplace and very large, requiring hardware support. Prediction · 2019 AI compute, hardware and data
Unresolvable Promise 28 Jul 2016 LeCun will teach his NYU deep learning course again in the Spring of 2017. Promise · 28 Jul 2016 Their own work or organisation
Resolved true Factual 28 Jul 2016 There have been a number of recent works putting graphical models on top of ConvNets and training end to end. Factual · 28 Jul 2016 AI capabilities
Unresolvable Factual 28 Jul 2016 There have been many proposals to combine neural nets and structured prediction, going back to the early 1990s. Factual · 28 Jul 2016 AI capabilities
Mostly true Factual 28 Jul 2016 The check reading system LeCun and a colleague built at Bell Labs in the early 1990s used structured prediction on top of convolutional nets, called Graph Transformer Networks. Factual · 28 Jul 2016 AI capabilities
Unresolvable Factual 28 Jul 2016 Videos of LeCun's 2015 NYU deep learning course lectures were taken down for legal reasons, while the slides remain available. Factual · 28 Jul 2016 Their own work or organisation
Resolved true Factual 28 Jul 2016 Yann LeCun gave a recent series of 8 lectures on deep learning at Collège de France in Paris. Factual · 28 Jul 2016 Their own work or organisation
Unresolvable Factual 28 Jul 2016 LeCun's Collège de France lectures were taught in French and later dubbed in English. Factual · 28 Jul 2016 Their own work or organisation
Unresolvable Factual 27 May 2015 Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio. Factual · 27 May 2015 AI capabilities
Resolved true Factual 27 May 2015 Deep learning methods have dramatically improved the state of the art in speech recognition, visual object recognition and object detection. Factual · 27 May 2015 AI capabilities
Mixed Factual 27 May 2015 Deep learning methods have dramatically improved the state of the art in many other domains such as drug discovery and genomics. Factual · 27 May 2015 AI capabilities
Mostly true Factual 27 May 2015 Deep learning uses the backpropagation algorithm to indicate how a machine should change the internal parameters that compute each layer's representation from the previous layer's. Factual · 27 May 2015 AI capabilities
Unresolvable Factual 27 May 2015 The authors of the paper declare no competing financial interests. Factual · 27 May 2015 Science and research methods
Mostly true Factual 27 May 2015 The acknowledgements, dated 2015, state that the authors Y.L. and Y.B. are fellows of the Canadian Institute For Advanced Research (CIFAR). Factual · 27 May 2015 Field building and careers
Unresolvable Prediction 18 Feb 2015 Self-driving cars will use convolutional neural networks. Prediction · 18 Feb 2015 AI capabilities
Mostly true Factual 18 Feb 2015 Deep Learning systems augmented with a memory module are starting to show impressive results in natural language processing. Factual · 18 Feb 2015 AI capabilities
Unresolvable Prediction 18 Feb 2015 A form of common sense could be acquired by machines through predictive unsupervised learning. Prediction · 18 Feb 2015 AI capabilities
Unresolvable Factual 18 Feb 2015 Yann LeCun has been obsessed with integrating learning end to end so machines learn good representations for over 30 years. Factual · 18 Feb 2015 Their own work or organisation
Mostly true Factual 18 Feb 2015 The brain is much better than current artificial models at unsupervised learning. Factual · 18 Feb 2015 AI capabilities
Mostly true Factual 18 Feb 2015 Shallow learning systems have one or two layers, while deep learning systems typically have five to 20 layers. Factual · 18 Feb 2015 AI capabilities
Mostly true Factual 18 Feb 2015 The system would probably not be more accurate, but it can potentially recognize people among hundreds of millions. Factual · 18 Feb 2015 AI capabilities
Unresolvable Prediction 18 Feb 2015 Deep Learning will be part of the solution for generalized intelligence. Prediction · 18 Feb 2015 AI timelines and forecasting
Mostly true Factual 18 Feb 2015 A typical Deep Learning system is like a box with 500 million knobs, 1,000 light bulbs and 10 million training images. Factual · 18 Feb 2015 AI compute, hardware and data
Mostly true Prediction 18 Feb 2015 Natural language understanding will be the next frontier for Deep Learning. Prediction · 18 Feb 2015 AI timelines and forecasting
Mostly true Factual 18 Feb 2015 A trained classifier may also correctly classify cars and dogs it has never seen before. Factual · 18 Feb 2015 AI capabilities
Unresolvable Prediction 18 Feb 2015 Deep Learning is part of the solution for machines representing relational and factual knowledge, but not the whole answer. Prediction · 18 Feb 2015 AI capabilities
Unresolvable Prediction 18 Feb 2015 AI hype could easily lead to another AI "winter cycle". Prediction · 18 Feb 2015 AI timelines and forecasting
Unresolvable Factual 15 May 2014 Theoretical results suggest that learning good parameters for a small neural network can be as computationally hard as breaking RSA. Factual · 15 May 2014 Philosophy, ethics and society
Mostly true Factual 15 May 2014 The notions of partition functions, entropy, free energy and variational methods prevalent in the graphical models literature come from statistical physics. Factual · 15 May 2014 Science and research methods
Unresolvable Factual 15 May 2014 The limitations the questioner pointed out concern not just backprop but all learning algorithms that use gradient-based optimization. Factual · 15 May 2014 AI capabilities
Mostly true Factual 15 May 2014 The first commercially viable check reading system, deployed by AT&T/NCR in 1996, used a ConvNet rather than an SVM. Factual · 15 May 2014 AI capabilities
Mostly false Factual 15 May 2014 Convolutional nets were always the record holder on MNIST, and SVMs never really managed to beat them there. Factual · 15 May 2014 AI capabilities
Unresolvable Prediction 15 May 2014 Researchers will eventually figure out how to merge ConvNets and DeconvNets into a feed-forward plus feed-back system trainable supervised or unsupervised. Prediction · 15 May 2014 AI capabilities
Mostly true Factual 15 May 2014 No one uses generalization bounds to do model selection; practitioners use cross-validation. Factual · 15 May 2014 Science and research methods
Unresolvable Prediction 15 May 2014 Now or in the near future, most of the world's knowledge will be extracted by machines and reside in machines. Prediction · 15 May 2014 AI timelines and forecasting
Unresolvable Prediction 15 May 2014 As more data becomes available, machine learning methods will increasingly rely on learning. Prediction · 15 May 2014 AI capabilities
Mostly true Factual 15 May 2014 Speech is a domain with access to ridiculously large amounts of data and a very large number of categories. Factual · 15 May 2014 AI capabilities
Unresolvable Prediction 15 May 2014 Natural language processing is the next frontier for deep learning. Prediction · 15 May 2014 AI timelines and forecasting
Mostly true Factual 15 May 2014 Facebook AI Research uses Torch7 for many projects, as do Deep Mind and several groups at Google. Factual · 15 May 2014 Technology
Mostly true Prediction 2010 In the next decade, a major challenge of computer vision, machine learning and AI will be devising methods that learn good feature hierarchies from unlabeled and labeled data in an integrated fashion. Prediction · 2010 AI capabilities
Unresolvable Prediction 2010 The authors expected to see many new embedded vision systems based on ConvNets in the next few years. Prediction · 2010 AI compute, hardware and data
Resolved true Factual 2010 Each stage of a ConvNet is composed of a filter bank, some non-linearities, and feature pooling layers. Factual · 2010 AI capabilities
Resolved true Factual 2010 The idea behind multi-stage Hubel-Wiesel architectures is rooted in Hubel and Wiesel's 1962 work on the cat's primary visual cortex. Factual · 2010 Science and research methods
Mostly true Factual 2010 A face detection application on the CNP system was demonstrated achieving 10 fps on VGA images. Factual · 2010 AI compute, hardware and data
Mixed Factual 2010 ConvNets have been successfully deployed in many commercial applications from OCR to video surveillance, but require large amounts of labeled training samples. Factual · 2010 AI capabilities
Mostly true Factual 2010 Convolutional Networks are a biologically inspired trainable architecture that can learn invariant features. Factual · 2010 AI capabilities
Unresolvable Factual 2010 A big advantage of ConvNets for detection is their computational efficiency. Factual · 2010 AI capabilities
Resolved true Factual 2010 In many problems labeled data is scarce whereas unlabeled data is abundant. Factual · 2010 AI capabilities
Mostly true Factual 2006 The hinge loss with G_W(X)=W^T X and a ||W||^2 regularizer gives the linear support vector machine. Factual · 2006 AI capabilities
Mostly true Factual 2006 The hinge loss systematically chooses the most offending incorrect answer as the one whose energy is pulled up. Factual · 2006 AI capabilities
Resolved true Factual 2006 The generalized perceptron loss is always positive because its second term is a lower bound on its first term. Factual · 2006 Science and research methods
Mostly true Factual 2006 Support vector Markov models, max-margin Markov networks and conditional random fields are all sequence modeling systems that use linearly parameterized energy factors. Factual · 2006 AI capabilities
Resolved true Factual 2006 Inference algorithms in energy-based learning often give only approximate answers or do not guarantee finding the global energy minimum. Factual · 2006 AI capabilities
Unresolvable Factual 2006 The LVQ2 loss is a continuous approximation of the number of classification errors. Factual · 2006 AI capabilities
Resolved true Factual 2006 Each NLL iteration involves considerably more work because pushing up every incorrect answer is expensive when the second term's derivative has no analytical form. Factual · 2006 AI capabilities
Unresolvable Factual 2006 There exists at least one point in HP1 at which the loss is less than at all points in HP2. Factual · 2006 AI capabilities
Back to the list

Yann LeCun

Open its page

Prediction AI timelines and forecasting

“The next frontier for Deep Learning is natural language understanding.”

Yann LeCun, Facebook AI director, interviewed 18 February 2015 Facebook AI Director Yann LeCun on His Quest to Unleash Deep Learning (IEEE Spectrum)

Natural language understanding will be the next frontier for Deep Learning.

Mostly true True in substance, wrong in a detail.

What would count as true or false

Judge over the years after February 2015 (about 2015–2020). True: natural language understanding became the main area of major new deep-learning advances and research attention in that period. Signs include landmark work (attention and Transformers, BERT, GPT-series language models), benchmarks such as GLUE and SuperGLUE, and contemporary reviews describing NLP or NLU as where deep learning made its next big breakthroughs. Mostly true: NLU was one of the leading frontiers, roughly equal with another area. Mixed: NLU advanced substantially, but other areas (for example deep reinforcement learning or image generation) were seen as at least as much the 'next frontier'. Mostly false: NLU progress was secondary, and another area was clearly the next frontier. False: deep learning made little progress in, or paid little attention to, NLU in that period.

Evidence

  1. Last Words: Computational Linguistics and Deep Learning - Nautilus Nautilus (reprinting a Computational Linguistics "Last Words" column) · December 2015
    “So far, problems in higher-level language processing have not seen the dramatic error rate reductions from deep learning that have been seen in speech recognition and in object recognition in vision.”
  2. Last Words: Computational Linguistics and Deep Learning - Nautilus Nautilus (reprinting a Computational Linguistics "Last Words" column) · December 2015
    “The next big step for Deep Learning is natural language understanding, which aims to give machines the power to understand not just individual words but entire sentences and paragraphs.”
  3. Attention Is All You Need arXiv (Google authors)
    “We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely.”
  4. Mastering the game of Go without human knowledge - Nature Nature (DeepMind authors) · 19 October 2017
    “Here we introduce an algorithm based solely on reinforcement learning, without human data, guidance or domain knowledge beyond game rules.”
  5. Last Words: Computational Linguistics and Deep Learning - Nautilus Nautilus · 16 March 2017
    “The next big step for Deep Learning is natural language understanding, which aims to give machines the power to understand not just individual words but entire sentences and paragraphs.”
  6. Last Words: Computational Linguistics and Deep Learning - Nautilus Nautilus · 16 March 2017
    “Deep Learning waves have lapped at the shores of computational linguistics for several years now, but 2015 seems like the year when the full force of the tsunami hit the major Natural Language Processing (NLP) conferences.”
  7. Last Words: Computational Linguistics and Deep Learning - Nautilus Nautilus · 16 March 2017
    “So far, problems in higher-level language processing have not seen the dramatic error rate reductions from deep learning that have been seen in speech recognition and in object recognition in vision.”
  8. Attention Is All You Need arXiv (Google authors)
    “Our model achieves 28.4 BLEU on the WMT 2014 English-to-German translation task, improving over the existing best results, including ensembles, by over 2 BLEU.”
  9. The state of AI in 2019: Breakthroughs in machine learning, natural language processing, games, and knowledge graphs - ZDNET ZDNET · 8 July 2019
    “As a demonstration of how quickly progress is being made in NLP, they go on to add, the state-of-the art has increased from a GLUE score of 69 to 88 over 13 months. The human baseline level is 87. Progress was so much faster than anticipated that a new benchmark SuperGLUE has already been introduced.”
  10. The state of AI in 2019: Breakthroughs in machine learning, natural language processing, games, and knowledge graphs - ZDNET ZDNET · 8 July 2019
    “Key areas covered are reinforcement learning, applications in games and future directions, natural language processing breakthroughs, deep learning in medicine, and AutoML.”
  11. BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding arXiv (ar5iv rendering of Devlin et al., Google) · 3 April 2018
    “It obtains new state-of-the-art results on eleven natural language processing tasks, including pushing the GLUE score to 80.5% (7.7% point absolute improvement), MultiNLI accuracy to 86.7% (4.6% absolute improvement), SQuAD v1.1 question answering Test F1 to 93.2 (1.5 point absolute improvement) and SQuAD v2.0 Test F1 to 83.1 (5.1 point absolute improvement).”
  12. NeurIPS Poster Language Models are Few-Shot Learners NeurIPS · 1 October 2020
    “We demonstrate that scaling up language models greatly improves task-agnostic, few-shot performance, sometimes even becoming competitive with prior state-of-the-art fine-tuning approaches.”

Reasoning

The evidence includes the landmark language work the criteria name: the Transformer (E3, E8), BERT's state-of-the-art results on GLUE, MultiNLI and SQuAD (E11), and the NeurIPS 2020 paper "Language Models are Few-Shot Learners", which reports that scaling up language models greatly improves few-shot performance (E12). ZDNET in 2019 reported that GLUE progress was so much faster than anticipated that a new benchmark, SuperGLUE, had already been introduced (E9), and a column said 2015 seemed to be the year when 'the full force of the tsunami' of deep learning hit the major NLP conferences (E6). However, the same 2019 ZDNET review lists NLP breakthroughs alongside reinforcement learning and games (E10), and AlphaGo Zero (E4) was a major reinforcement-learning landmark. So the evidence shows NLU as one of the leading frontiers, but it does not show that NLU clearly dominated every other area.

Cite this claim

Yann LeCun, “Natural language understanding will be the next frontier for Deep Learning.” (said 18 February 2015). Verdict: mostly true. Reliability API, claim c_a3663ae32bf0, method 0.1.0+c876030c. https://epistacklabs.com/claims/c_a3663ae32bf0

Claims without a verdict

Is a verdict here wrong, or do you speak for Yann LeCun? Reply on the claim's own page, with the form at its foot (how replies work). Verdicts are made automatically; how verdicts and scores are made.