Wassim (Wes) Bouaziz
@_Vassim
AI Security PhD student @MetaAI and @Polytechnique Previously @ENS_ULM @ENS_ParisSaclay I confront equations and inequalities💡
Why does Meta open-source its models? I talked about it with @kawecki_maciej looking at Dino, our computer vision model with applications in forest mapping, medical research, agriculture and more. Open-source boosts AI access, transparency, and safety. youtube.com/watch?v=eNGafi…
🚀 We are happy to organize the BERT²S workshop @NeurIPSConf 2025 on Recent Advances in Time Series Foundation Models. 🌐 berts-workshop.github.io 📜Submit by August 22 🎓Speakers and panelists: @ChenghaoLiu15 Mingsheng Long @zoe_piran @danielle_maddix @atalwalkar @qingsongedu
heading to @icmlconf #ICML2025 next week! come say hi & i'd love to learn about your work :) i'll present this paper (arxiv.org/abs/2503.17514) on the pitfalls of training set inclusion in LLMs, Thursday 11am here are my talk slides to flip through: ai.stanford.edu/~kzliu/files/m…
An LLM generates an article verbatim—did it “train on” the article? It’s complicated: under n-gram definitions of train-set inclusion, LLMs can complete “unseen” texts—both after data deletion and adding “gibberish” data. Our results impact unlearning, MIAs & data transparency🧵
Our research on embodied AI agents that can perceive, learn, act and interact in the virtual and physical worlds. #metaAI #AIAgent #embodied #worldmodel #superintelligemce arxiv.org/abs/2506.22355
❓How to balance negative and positive rewards in off-policy RL❓ In Asymmetric REINFORCE for off-Policy RL, we show that giving less weight to negative rewards is enough to stabilize off-policy RL training for LLMs! 💪 (1/8) Paper: arxiv.org/abs/2506.20520
Here is the recording with the slides for those interested! 🎤 youtu.be/UONvP1TL0-g?fe… 📊drive.google.com/file/d/14ZIopS… 📑arxiv.org/pdf/2410.02724 @Cohere_Labs @Cohere_Labs
Our ML Theory group is looking forward to welcoming @AmbroiseOdonnat next week on Thursday, June 19th for a session on "Large Language Models as Markov Chains"
We present an Autoregressive U-Net that incorporates tokenization inside the model, pooling raw bytes into words then word-groups. AU-Net focuses most of its compute on building latent vectors that correspond to larger units of meaning. Joint work with @byoubii 1/8
DINOv2 meets text at #CVPR 2025! Why choose between high-quality DINO features and CLIP-style vision-language alignment? Pick both with dino.txt 🦖📖 We align frozen DINOv2 features with text captions, obtaining both image-level and patch-level alignment at a minimal cost. [1/N]
🚀To know more about LLM as Markov Chains, join in on June 19th at 6 pm CET (Paris time)!!😀 Huge thanks to @itsmaddox_j and @cohere @Cohere_Labs for the invitation 🤗 Paper: arxiv.org/pdf/2410.02724 meeting: cohere.com/events/Cohere-…
Our ML Theory group is looking forward to welcoming @AmbroiseOdonnat next week on Thursday, June 19th for a session on "Large Language Models as Markov Chains"
🚨 Your RL only improves 𝗽𝗮𝘀𝘀@𝟭, not 𝗽𝗮𝘀𝘀@𝗸? 🚨 That’s not a bug — it’s a 𝗳𝗲𝗮𝘁𝘂𝗿𝗲 𝗼𝗳 𝘁𝗵𝗲 𝗼𝗯𝗷𝗲𝗰𝘁𝗶𝘃𝗲 you’re optimizing. You get what you optimize for. If you want better pass@k, you need to optimize for pass@k at training time. 🧵 How?
I'm in Singapore this week to present Data Taggants at ICLR 😃 I'll be at Friday 25th morning's poster session :) Find me in spot #565 from 10h to 12h30!
Want to know if a ML model was trained on your dataset with 1 API call? See you in conferences 🙌 Excited to share that our paper Data Taggants for image data was accepted at ICLR 2025 🎉 Our follow-up on audio data, was accepted at ICASSP 2025! 🎉 Check out the details below 👇