Goto

Collaborating Authors

 Industry

One-Shot Safety Alignment for Large Language Models via Optimal Dualization

Neural Information Processing Systems

Ideally, we would like methods that train LMs only once ( i.e., one-shot) with a fixed objective, as in

bf0857cb9a41c73639f028a80301cdf0-Paper-Conference.pdf

996bef37d8a638f37bdfcac2789e835d-Paper-Conference.pdf

Aiming towards the minimizers: fast convergence of SGD for overparameterized problems

Neural Information Processing Systems

Recent advances in machine learning and artificial intelligence have relied on fitting highly overparam-eterized models, notably deep neural networks, to observed data; e.g.

9961e42624a6c083279303767c73269d-Paper-Conference.pdf

Keypoint-Augmented Self-Supervised Learning for Medical Image Segmentation with Limited Annotation

Neural Information Processing Systems

Pretraining CNN models (i.e., UNet) through self-supervision has become a powerful approach to facilitate medical image segmentation under low annotation regimes.

9952369f49ecc064d169fe6612cbf204-Paper-Conference.pdf

99261adc8a6356b38bcf999bba9a26dc-Paper-Conference.pdf

Reward-Directed Conditional Diffusion: Provable Distribution Estimation and Reward Improvement

Neural Information Processing Systems

We explore the methodology and theory of reward-directed generation via conditional diffusion models.

Reward-Directed Conditional Diffusion: Provable Distribution Estimation and Reward Improvement

Neural Information Processing Systems

We explore the methodology and theory of reward-directed generation via conditional diffusion models.