Les demandes de prise en charge de missions par le GdR IASIS doivent parvenir à la gestionnaire du GdR avant le 25 septembre. Les modèles génératifs ont connu de récentes avancées spectaculaires, au p...

Multi-reward conditioned text-to-image diffusion (ICML 2026)

The default paradigm of post-training text-to-image generators includes post-hoc selection of generated images, and subsequent training with one reward model to align the generator to the reward, typi...

An interactive companion to 'The FID Lottery'. Every reported FID is the outcome of two lotteries — we measure how much they move the number.

Geometry is invariant to viewpoint, which makes any collection of images a redundant encoding of a single 3D state. Existing feed-forward reconstruction models fail to exploit this: per-view methods e...

KE:SAI is a Franco-German non-profit open science lab for scalable autonomous intelligence.

The default paradigm of post-training text-to-image generators includes post-hoc selection of generated images, and subsequent training with one reward model to align the generator to the reward, typi...

Back into Plato's Cave: Examining Cross-modal Representational Convergence at Scale

Our mission is to build and democratize artificial general intelligence through open science.

This paper introduces the Polynomial Mixer (PoM), a novel token mixing mechanism with linear complexity that serves as a drop-in replacement for self-attention. PoM aggregates input tokens into a comp...

Monocular novel-view synthesis has long required multi-view image pairs for supervision, limiting training data scale and diversity. We argue it is not necessary: one view is enough. We present OVIE, ...

YouTube video by Nicolas Dufour

Train once, align many rewards. MIRO achieves 19× faster convergence and 370× less compute than FLUX while reaching GenEval score of 75. Controllable trade-offs at inference time.