Back to Homepage

Samy Bengio

Senior Director, AI and Machine Learning Research, Apple Inc
Adjunct Professor, EPFL
Apple Park, Cupertino, CA 95014, USA bengio@apple.com http://bengio.abracadoudou.com/ French and English Canadian

Executive Summary

Senior Director of AI and Machine Learning Research at Apple and Adjunct Professor at EPFL. Over 30 years of foundational and applied research experience across neural networks, deep learning architectures, multimodal learning, large language models, and machine learning systems. Co-creator of seminal open-source machine learning frameworks including Torch, SVMTorch, and contributor to TensorFlow; executive sponsor of MLX for Apple Silicon. General Chair of NeurIPS 2018, Program Chair of NIPS 2017 and ICLR 2015–2016. Member of the NeurIPS Board (2019–2025) and NeurIPS Scientific Advisory Board (2026–). Author of over 320 publications with over 130,000 citations.

Education

Ph.D. in Computer Science 1989–1993
Université de Montréal, Canada
Thesis: Optimization of a Parametric Learning Rule for Neural Networks
(Optimisation d'une règle d'apprentissage pour réseaux de neurones artificiels)
M.Sc. in Computer Science 1986–1989
Université de Montréal, Canada
Thesis: Integration of Traditional and Intelligent Tutoring Systems
(Intégration des systèmes tutoriels traditionnels et des systèmes tutoriels intelligents)
B.Sc. in Computer Science 1983–1986
Université de Montréal, Canada

Employment History

Adjunct Professor

2024–Present
EPFL (École Polytechnique Fédérale de Lausanne), Switzerland

Academic research and doctoral supervision in machine learning foundations and reasoning.

Senior Director, AI and Machine Learning Research

2021–Present
Apple Inc, Cupertino, CA, USA

Leading foundational and curiosity-driven machine learning research across deep learning theory, autoregressive reasoning, multimodal intelligence, and efficient ML systems (MLX).

Distinguished Research Scientist

2007–2021
Google Inc (Google Brain & Research), Mountain View, CA, USA

Research on machine learning algorithms and deep architectures for large-scale ranking, image/video understanding, multimodal perception, and speech. Co-author of seminal deep learning works (ICLR'17 Best Paper on Generalization, Show and Tell image captioning).

Senior Researcher in Machine Learning

1999–2007
IDIAP Research Institute, Martigny, Switzerland

Research on statistical machine learning, kernel methods, sequence processing, multimodal interaction, and biometrics. Supervised 17 Ph.D. students and postdocs. Principal investigator on major Swiss and European research grants.

Research Director

1997–1999
Microcell Labs, Montréal, Canada

Directed a research team applying learning algorithms to large-scale data warehouses for telecommunications: fraud detection, customer churn prediction, and traffic forecasting.

Researcher

1996–1997
CIRANO (Centre Interuniversitaire de Recherche en Analyse des Organisations), Montréal, Canada

Developed statistical and machine learning models for financial market prediction, portfolio selection, and autonomous auction agents.

Postdoctoral Fellow

1995–1996
INRS-Télécommunications (Institut National de la Recherche Scientifique), Montréal, Canada

Developed Asynchronous Input/Output Hidden Markov Models (connectionist recurrent mixture of experts) for non-stationary temporal speech recognition.

Postdoctoral Fellow

1994–1995
CNET (Centre National d'Études des Télécommunications), France Télécom, Lannion, France

Time series prediction with neural networks and connectionist mixture models for physical and telecommunications data.

Research Assistant

1993
CRT (Centre de Recherche sur les Transports), Université de Montréal, Canada

Applied Genetic Algorithms and Tabu Search to combinatorial optimization problems (Vehicle Routing Problem with Time Windows).

System Administrator & Research Assistant

1986–1993
Department of Computer Science, Université de Montréal, Canada

Part-time system administrator for multimedia and intelligent tutoring laboratories.

Research Projects Funded & Managed

Swiss National Science Foundation (SNSF) Projects

  • IM2.ACP, IM2.MI, IM2.BMI, and IM2.MPR — Interactive Multimodal Information Management (4 PhDs, 2 postdocs)
  • MULTI — Multimodal Interaction and Multimedia Data Mining (several PhDs)
  • KerSpeech — Kernel Methods for Speech and Video Sequence Analysis (1 PhD)
  • ADASEQ — Ensemble Methods for Sequence Processing (1 PhD)
  • CARTANN — Cartography by Artificial Neural Networks (1 PhD completed)
  • Divide and Learn I and II — Mixture Models for Large Datasets (3 PhDs, 1 thesis completed)
  • GLAD — Use of Boolean Methods for Classification (1 PhD completed)
  • KERNEL — Kernel Methods for Sequence Processing (1 PhD)
  • SCRIPT — Cursive Handwriting Recognition (1 PhD completed)

European Framework Projects

  • PASCAL — Pattern Analysis, Statistical Modelling and Computational Learning (EU FP6 Network of Excellence; Member of Steering Committee)
  • BANCA — Biometric Access Control for Networked and e-Commerce Applications (EU FP5, 2 researchers)
  • EDAM — Environmental Data Mining: Machine Learning Algorithms and Statistical Tools for Monitoring and Forecasting (INTAS Foundation, 1 invited researcher)
  • LAVA — Learning for Adaptable Visual Assistants (1 postdoc, 2 PhDs)
  • COST-275 — Biometric-Based Recognition of People over the Internet (1 PhD)

Scientific Leadership & Duties

Advisory Boards & Governance

  • NeurIPS Scientific Advisory Board (2026–Present)
  • NeurIPS Board Member (2019–2025)
  • ICLR Board Member (2018–2020)
  • University of Montreal International Advisory Committee to the Rector (2022–Present)
  • Simons Institute for the Theory of Computing Advisory Board (2018–2020)
  • NSF ID4 (Institute for Data Driven Dynamical Design) Advisory Board (2022–Present)
  • MLMI Advisory Board (2009)

Editorial Leadership

  • Action Editor, Journal of Machine Learning Research (JMLR) (2012–Present)
  • Editorial Board Member, Journal of Machine Learning Research (JMLR) (2009–2012)
  • Editorial Board Member, Machine Learning Journal (MLJ) (2010–2020)
  • Associate Editor, Journal of Computational Statistics (2002–2011)
  • Associate Editor, IEEE Journal of Selected Topics in Signal Processing (2009)

Conference Leadership

  • General Chair, NeurIPS: Neural Information Processing Systems (2018)
  • Programme Chair, NIPS: Neural Information Processing Systems (2017)
  • Programme Chair, ICLR: International Conference on Learning Representations (2015, 2016)
  • Ethics Review Co-Chair, NeurIPS (2021)
  • Demonstrations Chair, NIPS (2011)

Workshop & Special Conference Organization

  • Founder & Organizer, BayLearn: Bay Area Machine Learning Symposium (2012–2016)
  • Co-organizer, BigVision: NIPS Workshop on Big Data for Computer Vision (2012)
  • Co-organizer, BigVision: CVPR Workshop on Big Data for Computer Vision (2014, 2015)
  • Co-organizer, WebVision: ECCV Workshop on Computer Vision for the Web (2012)
  • Co-organizer, MLMI: Machine Learning for Multimodal Interaction (2004, 2005, 2006)
  • Co-organizer, NIPS Workshop on Efficient Machine Learning (2007)
  • Co-organizer, NIPS Workshop on Learning to Compare Examples (2006)
  • Co-organizer, NIPS Workshop on Multimodal Signal Processing (2004)

Senior Area Chair & Area Chair

  • Senior Area Chair, ICLR (2022, 2023, 2024)
  • Senior Area Chair, ICML (2019, 2021, 2022, 2023, 2024)
  • Senior Area Chair, NeurIPS (2020, 2022, 2023, 2024)
  • Area Chair, ICLR (2019, 2021)
  • Area Chair, ICML (2009, 2012, 2015, 2016, 2020)
  • Area Chair, NeurIPS / NIPS (2003, 2006, 2012, 2014, 2015)
  • Senior Program Committee, IJCAI (2011)
  • Area Chair, ECML-PKDD (2010)
  • Area Chair, IEEE NNSP (2002)

Ph.D. Thesis Juries & Academic Committees

  • Member of doctoral examination juries across top international institutions:
  • Françoise Fessant, Université de Rennes (1995)
  • Sébastien Marcel, Université de Rennes (2000)
  • Miguel Moreira, EPFL Lausanne (2001)
  • Pierre-Edouard Sottas, EPFL Lausanne (2002)
  • Nicolas Gilardi, Université de Lausanne (2002)
Show 34 more
  • Todd Stephenson, EPFL Lausanne (2003)
  • Katrin Weber, EPFL Lausanne (2003)
  • Torsten Butz, EPFL Lausanne (2003)
  • Liva Ralaivola, Université de Paris 6 (2003)
  • Ronan Collobert, Université de Paris 6 (2004)
  • Vlad Popovici, EPFL Lausanne (2004)
  • Mathieu Ben, Université de Rennes (2004)
  • Fabien Cardinaux, EPFL Lausanne (2005)
  • Serghei Kosinov, Université de Genève (2005)
  • Jean-Julien Aucouturier, Université de Paris 6 (2006)
  • Sylvain Ferrandiz, Université de Caen (2006)
  • Silvia Chiappa, EPFL Lausanne (2006)
  • Norman Poh, EPFL Lausanne (2006)
  • Alexei Pozdnoukhov, EPFL Lausanne (2006)
  • Mikaela Keller, EPFL Lausanne (2006)
  • Christos Dimitrakakis, EPFL Lausanne (2006)
  • Dong Zhang, EPFL Lausanne (2006)
  • Gianluca Monaci, EPFL Lausanne (2007)
  • Ivana Arsic, EPFL Lausanne (2007)
  • David Grangier, EPFL Lausanne (2008)
  • Jean-François Paiement, EPFL Lausanne (2008)
  • Marie Szafranski, Université de Technologie de Compiègne (2008)
  • Mihai Gurban, EPFL Lausanne (2008)
  • Jie Luo, EPFL Lausanne (2011)
  • Majid Yazdani, EPFL Lausanne (2013)
  • Pierre-Michel Bousquet, Université d'Avignon (2014)
  • Seungyeon Kim, Georgia Tech (2015)
  • Dimitri Palaz, EPFL Lausanne (2016)
  • Hieu Pham, Carnegie Mellon University (2021)
  • Paul Best, Université de Toulon (2022)
  • Sotirios Anagnostidis, ETH Zürich (2025)
  • Youssef Allouah, EPFL Lausanne (2025)
  • Hervé Glotin, Université Sud Toulon-Var (2007)
  • Vincent Lemaire, Université de Paris-Sud (2008)

Teaching Experience

Course IC-49: Statistical Machine Learning from Data 2005–2006
EPFL — Doctoral Program in Computer, Communication and Information Sciences
Comprehensive graduate lectures on learning theory, SVMs, graphical models, and neural networks.
Lecture Notes & Slides
Advanced Lectures on Statistical Machine Learning 2002–2003
IDIAP Research Institute, Martigny, Switzerland
Lectures on kernel methods, ensemble learning, and structural risk minimization.
Lecture Materials
Graduate Teaching Replacement (M.Sc. / Ph.D.) 1992–1993, 1996
Université de Montréal, Canada
Lectures in Artificial Neural Networks, Genetic Algorithms, and Hidden Markov Models.
Sprint School on Learning Algorithms 1996
Université de Montréal, Canada
Intensive workshop on Hidden Markov Models and Input/Output HMMs.

Publications

320 Publications Archive

Complete Research Bibliography (1987–2026)

All 320 publications spanning peer-reviewed journal articles, international conference proceedings, book chapters, and research reports with direct PDF downloads, web links, and interactive abstracts.

Software & Systems Development

MLX Framework
Open-source machine learning framework natively designed for Apple Silicon with unified memory architecture and lazy evaluation.
TensorFlow
Core contributions to large-scale distributed machine learning infrastructure at Google.
Torch Library
Pioneering C++/Lua machine learning library that served as the foundational precursor to PyTorch.
SVMTorch
Large-scale Support Vector Machine package for classification and regression tasks.
Languages
Python, C, C++, Objective-C, Swift, Java, Lisp, Matlab, Mathematica, UNIX shell.
Systems
Decades of deep multi-platform system administration and high-performance computing experience (UNIX, Linux, macOS).

References

Academic and professional references are available upon request.