AI People Wiki

Builders & open source

Quoc Le

Co-author of the sequence-to-sequence, GNMT, and neural architecture search papers at Google

Quoc V. Le co-authored Sequence to Sequence Learning with Neural Networks (2014) with Ilya Sutskever and Oriol Vinyals, all listed at Google. With a Google Brain affiliation, Le co-authored Neural Architecture Search with Reinforcement Learning (2016) with Barret Zoph and the 2017 paper that introduced the sparsely-gated mixture-of-experts layer, and Le is also a co-author of the 2016 paper presenting GNMT, Google's neural machine translation system.1234

At a glance

Current role
None confirmed in the sources
Education
Stanford
Work listed
4
Sources cited
10
Profile reviewed

Representation learning

Quoc V. Le co-authored Building high-level features using large scale unsupervised learning (2011), which trained a large network on unlabeled images to ask whether class-specific detectors, such as a face detector, can emerge without labels. With Tomas Mikolov, Le wrote Distributed Representations of Sentences and Documents (2014), which learns fixed-length vector representations of text as an alternative to bag-of-words features.56

Sequence models and translation

In 2014 Le co-authored Sequence to Sequence Learning with Neural Networks with Ilya Sutskever and Oriol Vinyals, which maps input sequences to output sequences with multilayer LSTMs. Le was also a co-author of the 2016 paper presenting GNMT, Google's neural machine translation system, of the 2017 sparsely-gated mixture-of-experts layer, and of a 2017 large-scale study of neural machine translation architectures.1437

Automated design and language-model reasoning

With Barret Zoph, Le co-authored Neural Architecture Search with Reinforcement Learning (2016), and with Mingxing Tan wrote EfficientNet (2019), which balances network depth, width and resolution when scaling convolutional networks and uses architecture search to design the base model.28

Later papers co-authored by Le address large language models: Finetuned Language Models Are Zero-Shot Learners (2021), which showed that instruction tuning improves zero-shot performance and called the resulting model FLAN, and Chain-of-Thought Prompting Elicits Reasoning in Large Language Models (2022), which showed that prompting with intermediate reasoning steps improves performance on complex reasoning tasks.910

Work and contributions

  • Sequence to sequence learning Published research

    Co-author of Sequence to Sequence Learning with Neural Networks (2014), which maps an input sequence to an output sequence with multilayered LSTM networks.1

  • Neural architecture search Published research

    Co-author, with Barret Zoph, of the 2016 Google Brain paper that trains a recurrent network with reinforcement learning to generate neural network architectures.2

  • Google's Neural Machine Translation system (GNMT) Published research

    Co-author of the 2016 paper presenting GNMT, Google's neural machine translation system.4

  • Sparsely-gated mixture-of-experts layer Published research

    Co-author of Outrageously Large Neural Networks (2017), which applies conditional computation through a sparsely-gated mixture-of-experts layer.3

Timeline

  1. Co-authored Building high-level features using large scale unsupervised learning.5
  2. Released Distributed Representations of Sentences and Documents with Tomas Mikolov.6
  3. Co-authored Sequence to Sequence Learning with Neural Networks.1
  4. Co-authored the paper presenting Google's Neural Machine Translation system.4
  5. Co-authored Neural Architecture Search with Reinforcement Learning with Barret Zoph.2
  6. Co-authored the sparsely-gated mixture-of-experts layer paper.3
  7. Released EfficientNet with Mingxing Tan.8
  8. Co-authored the FLAN instruction-tuning paper.9
  9. Co-authored Chain-of-Thought Prompting Elicits Reasoning in Large Language Models.10

Documented roles

The sources do not confirm a current role; these are the roles they document.

  • Listed with Google and Google Brain affiliations on 2014–2017 papers, Google Brain1237

Companies in their history

Suggest a correction or removal

Profiles cover public professional work only, and each statement cites a source. If something is out of date, missing or wrong, tell us and link a source. People can also ask for their profile to be corrected or removed.

References

10 sources · profile reviewed Oct 21, 2026 · data updated Oct 21, 2026

  1. Sequence to Sequence Learning with Neural Networks (arXiv 1409.3215)arXiv · Paper · published Sep 10, 2014 · accessed Oct 21, 2026
  2. Neural Architecture Search with Reinforcement Learning (arXiv 1611.01578)arXiv · Paper · published Nov 5, 2016 · accessed Oct 21, 2026
  3. Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer (arXiv 1701.06538)arXiv · Paper · published Jan 23, 2017 · accessed Oct 21, 2026
  4. Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation (arXiv 1609.08144)arXiv · Paper · published Sep 26, 2016 · accessed Oct 21, 2026
  5. Building high-level features using large scale unsupervised learning (arXiv 1112.6209)arXiv · Paper · published Dec 29, 2011 · accessed Oct 21, 2026
  6. Distributed Representations of Sentences and Documents (arXiv 1405.4053)arXiv · Paper · published May 16, 2014 · accessed Oct 21, 2026
  7. Massive Exploration of Neural Machine Translation Architectures (arXiv 1703.03906)arXiv · Paper · published Mar 11, 2017 · accessed Oct 21, 2026
  8. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks (arXiv 1905.11946)arXiv · Paper · published May 28, 2019 · accessed Oct 21, 2026
  9. Finetuned Language Models Are Zero-Shot Learners (arXiv 2109.01652)arXiv · Paper · published Sep 3, 2021 · accessed Oct 21, 2026
  10. Chain-of-Thought Prompting Elicits Reasoning in Large Language Models (arXiv 2201.11903)arXiv · Paper · published Jan 28, 2022 · accessed Oct 21, 2026