Antonio Torralba

Generalized Representation for Generalized Dynamics Generation

MathNet: A Global Multimodal Benchmark for Mathematical Reasoning and Retrieval

Where is Motion From? Scalable Motion Attribution for Video Generation Models

DiffEM: Learning from Corrupted Data with Diffusion Models via Expectation Maximization

Seeing Through the PRISM: Compound & Controllable Restoration of Scientific Images

Ambient Dataloops: Generative Models for Dataset Refinement

Fairness Aware Reward Optimization

Dataset Distillation for Pre-Trained Self-Supervised Vision Models

Ambient Diffusion Omni: Training Good Models with Bad Data

Feed-Forward Bullet-Time Reconstruction of Dynamic Scenes from Monocular Videos

Automated Detection of Visual Attribute Reliance with a Self-Reflective Agent

Multiagent Finetuning: Self Improvement with Diverse Reasoning Chains

Skin, Muscles, and Bones in MultiSensory Simulation

LoRA vs Full Fine-tuning: An Illusion of Equivalence

Single-pass Adaptive Image Tokenization for Minimum Program Search

Adaptive Length Image Tokenization via Recurrent Allocation

LoRA vs Full Fine-tuning: An Illusion of Equivalence

Separating Knowledge and Perception with Procedural Data

Art-Free Generative Models: Exploring Art Creation Without Prior Artistic Knowledge