iiinova — where your digital self keeps evolving

Samuel Lavoie

@lavoiems.bsky.social

291 followers 109 following 13 posts

PhD candidate @Mila_quebec, @UMontreal. Ex: FAIR @AIatMeta. Learning representations, minimizing free energy, running.

Posts Replies Media

Reposted by Samuel Lavoie

Guillaume Lajoie @glajoie.bsky.social · 19/08/2025

Compositionality is a central desideratum for intelligent systems...but it's a fuzzy concept and difficult to quantify. In this blog post, lab member @ericelmoznino.bsky.social outlines ideas toward formalizing it & surveys recent work. A must-read for interested researchers in AI and Neuro

0215

Samuel Lavoie @lavoiems.bsky.social · 22/07/2025

🧵 Everyone is chasing new diffusion models—but what about the representations they model from? We introduce Discrete Latent Codes (DLCs): - Discrete representation for diffusion models - Uncond. gen. SOTA FID (1.59 on ImageNet) - Compositional generation - Integrates with LLM 🧱

153

Samuel Lavoie @lavoiems.bsky.social · 17/07/2025

The code and model weights for Llip are finally out! I hope you will find this model useful! Paper: arxiv.org/abs/2405.00740 Code: github.com/facebookrese... Models: - ViT-G: huggingface.co/lavoies/llip... - ViT-B: huggingface.co/lavoies/llip...

arxiv.org

Modeling Caption Diversity in Contrastive Vision-Language Pretraining

There are a thousand ways to caption an image. Contrastive Language Pretraining (CLIP) on the other hand, works by mapping an image and its caption to a single vector -- limiting how well CLIP-like mo...

061