🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Even more than that: as long as you are creative and rigorous, you can do amazing AI research without even knowing how to implement a transformer! It's the beauty of AI research: you can pick your preferred level of abstraction, and there is a need for people doing work at any

I gave a talk on Procedure Understanding from Egocentric Videos @ the Eyes of the Future Workshop- ECCV 2024 lnkd.in/dWCxctsX The talk was based on recent our works (you can see 3 of them at ECCV poster session). Thank you to M. Matteucci/F.Palermo for inviting me!

Antonino Furnari presenting Ego-Exo4D @ the Third Hands-on Egocentric Research Tutorial with Project Aria

📢 ⚡️Finally glad to share our preliminary assessments on CALM (Credit Assignment with Language Models). We use LLMs to decompose an RL 🤖 task into hierarchical subgoals 🎯 and ask them to verify their achievements.

BREAKING NEWS The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Physics to John J. Hopfield and Geoffrey E. Hinton “for foundational discoveries and inventions that enable machine learning with artificial neural networks.”

Mila's annual supervision request process opens on October 15 to receive MSc and PhD applications for Fall 2025 admission! Join our community! More information here mila.quebec/en/prospective…

How certain are you that your #DeepRL algorithms actually work? Maybe you tested them on one of the standard benchmark tasks like the beloved OpenAI Gym Hopper … or maybe you chose the DMC Hopper? Does this choice influence your conclusions? arxiv.org/abs/2410.08870

I had a fantastic time talking with Sam Charrington (The TWIML AI Podcast) about Orion Publishing' "Artificial Intelligence: 10 Things You Should Know" (geni.us/ArtificialInte…) and many exciting 2024 research papers (some of them from my teams) in the Open-Endedness community by outstanding

Life update! I have returned to FAIR (AI at Meta) 50% of my time where I'll be supporting Yoram Bachrach in building up the Multi-Agent Universal Intelligence (MAUI) team in London. Instead of playing catchup, MAUI's mission are methods which allow open-source and science to leapfrog!

Shouldn't whatever comes after "pre-training" be called just "training" instead of "post-training"?

Last week I successfully defended my PhD thesis titled "Building Intuitive Reinforcement Learning Algorithms". I am very grateful to anybody who contributed to this journey!

Proud to have been part of the team behind Meta Motivo, a truly groundbreaking foundation model for behavior. It’s the first of its kind, enabling you to instantly generate human-like behaviors for any reward function or goal. Make sure to check out the demo for yourself!

As I'm heading out of Vancouver, I'm also wrapping up my postdoc at FAIR. It was a fun week with my friends/collabs and I made many great connections. But, hey! I'm still looking for full-time positions. Reach out if you work on multimodal generation/understanding #NeurIPS2024

🤔 How to extract knowledge from LLMs to train better RL agents? 📚 Our new paper (with Qinqing Zheng Mikael Henaff Amy Zhang Aditya Grover ) studies LLM-driven rewards for NetHack! Paper: arxiv.org/abs/2410.23022 Code: github.com/facebookresear…

Excited to share our latest work ONI, which enables learning intrinsic rewards online without pre-collected data. We do this by annotating the agent's collected experience with an asynchronously hosted LLM server. Paper: arxiv.org/abs/2410.23022 Code: github.com/facebookresear…

What if you could just describe in words to an AI agent the skills it should learn? The future of AI is human-AI collaboration. You are now able to convert human knowledge about useful skills into the ability to learn them via RL and to compose them with codegen. Check it out!

Super excited to see MaestroMotif out into the world--the first hierarchical LLM agent that can solve open-ended compositional tasks requiring hundreds of steps 🚀🚀🚀 🤖 What can MaestroMotif do? - solve complex tasks by re-combining skills - adapt zero-shot to new instructions

The era of Hierarchical Agents has begun

See our work on training MoEs using reinforcement learning for molecule generation!