Facebook

VP & Chief AI Scientist at Meta

Scientific and technical leadership at Facebook AI Research, AI R&D strategy.

VP & Chief AI Scientist

Director of Facebook AI Research (FAIR), with research locations in Menlo Park, New York City, and Paris, focused on bringing about major advances in AI.

Professor at NYU on a part-time basis (since January 2014).

Cofounder and advisor of Elements Inc. (biometrics), and Museami (music technology)

Director of AI Research

New York University

Silver Professor of Computer Science, Courant Institute of Mathematical Sciences (since 2006)
Professor of Neural Science, NYU Center for Neural Science (since 2007)
Professor of Electrical and Computer Engineering, NYU School of Engineering (since 2011)
Professor of Data Science, NYU Center for Data Science (since 2013)
Founding Director of the NYU Center for Data Science (since 2013)

Research and teaching in machine learning, AI, data science, computer vision, robotics, computational neuroscience, and knowledge extraction from data,

Professor

Element Inc

Elements develops software-based biometric authentication technology.

Co-founder, Advisor

MuseAmi

Development of innovative software and hardware technology for, music production, music education, and musical entertainment. Products includes mobile apps such as Improvox, Hook'd and MusicPal.

Cofounder and Chief Scientist

YLC Consulting LLC

Research, Development, and Technology Consulting in software, machine learning, machine vision, robotics, data mining, data modeling, statistical data analysis, OCR, data compression, music technology and related topics .

Owner

NEC Research Institute

Research in machine learning and computer vision.

Fellow

AT&T Labs Research

Head of the Image Processing Research Department.

Directed research in machine learning, pattern recognition, image compression, video compression, animated characters. Department members included Yoshua Bengio, Leon Bottou, Eric Cosatto, Hans-Peter Graf, Patrick Haffner, Fu Jie Huang, Jörn Ostermann, Patrice Simard, Vladimir Vapnik,

Department Head

AT&T Bell Laboratories

Carried out research in machine learning, neural nets, optical character recognition, handwriting recognition, signature verification, integrated circuits for pattern recogntiion, and other topics.

research staff member

Research Staff Member

Research staff member

THALES - Laboratoire Central de Recherche

research scientist

University of Toronto

postdoc in the Geoff Hinton's lab in the Department of Computer Science.

Research Associate (postdoc)

Professor, researcher, and R&D manager with academic and industry experience in AI, machine learning, deep learning, computer vision, intelligent data analysis, data mining, data compression,  digital library systems, and robotics.

Specialties: research, technical consulting, scientific advising.

🥁 Llama3 is out 🥁
8B and 70B models available today.
8k context length.
Trained with 15 trillion tokens on a custom-built 24k GPU cluster.
Great performance on various benchmarks, with Llam3-8B doing better than Llama2-70B in some cases.
More versions are coming over the next few months.

https://lnkd.in/dNZx72FJ


Yann LeCun

That was fast!
And it's only the beginning...

Why is it important to fine-tune models for specific use-cases?

How can llama3 models be used in industry?

What would be the significance of having 10,000 llama3 variants?

Already almost 1,000 llama3 model variations have been shared publicly on HF (many more in private use at companies): https://lnkd.in/enMHktPX

Everyone should fine-tune their own models for their use-cases, languages, industry, infra constraints,... 

10,000 llama3 variants by the end of next week?

Video of the Ding-Shum Lecture I gave at Harvard's Center of Mathematical Sciences and Applications on 2024-03-28.

"Objective-Driven AI: Towards AI systems that can learn, remember, reason, and plan"
Abstract: How could machines learn as efficiently as humans and animals?
How could machines learn how the world works and acquire common sense?
How could machines learn to reason and plan?
Current AI architectures, such as Auto-Regressive Large Language Models fall short. I will propose a modular cognitive architecture that may constitute a path towards answering these questions. The centerpiece of the architecture is a predictive world model that allows the system to predict the consequences of its actions and to plan a sequence of actions that optimize a set of objectives. The objectives include guardrails that guarantee the system's controllability and safety. The world model employs a Hierarchical Joint Embedding Predictive Architecture (H-JEPA) trained with self-supervised learning. The JEPA learns abstract representations of the percepts that are simultaneously maximally informative and maximally predictable. 

Working paper: https://lnkd.in/g-h6cqRy
Slides: https://lnkd.in/eTq65_tF
Video: https://lnkd.in/efGkys4i

Giving a speech at the AI Dinner in Paris's Station F.
#raybanmeta


OpenDAC: an open dataset to train AI systems to help design compounds for the direct capture of CO2 from the air.

https://lnkd.in/dW83UtxD

Taplio

Yann LeCun's Linkedin Analytics

Yann LeCun

What is Yann talking about?

Yann LeCun's Best Posts (last 30 days)

Want to drive more opportunities from LinkedIn?

Taplio

.css-1jgreyo{background-image:linear-gradient(284deg, rgba(39,43,72,1) 0%, rgba(33,47,183,1) 100%);color:transparent;-webkit-background-clip:text;background-clip:text;}Yann LeCun's Linkedin Analytics

Yann LeCun

What is Yann talking about?

Yann LeCun's Best Posts .css-qqfgvy{font-size:var(--chakra-fontSizes-md);}(last 30 days)

Want to drive more opportunities from LinkedIn?

Yann LeCun's Linkedin Analytics

Yann LeCun's Best Posts (last 30 days)