We introduce a selector-extractor framework that extracts high-res features without ever seeing full high-res images to save compute.

WhAM: a transformer model unifying generation, acoustic translation and classification of sperm whale vocalizations

We presents Reward Dithering, a technique that enhances reinforcement learning in large language models by adding random perturbations to reward signals, improving training efficiency and convergence speed while maintaining performance.

We unify submodular/supermodular ratio problems and show general algorithms like SuperGreedy++ and min-norm point methods are surprisingly efficient and scalable, outperforming specialized methods in theory and practice