Inference-time scaling for diffusion models beyond scaling denoising steps

N Ma, S Tong, H Jia, H Hu, YC Su, M Zhang… - arxiv preprint arxiv …, 2025 - arxiv.org
Generative models have made significant impacts across various domains, largely due to
their ability to scale during training by increasing data, computational resources, and model …

LLMs are highly-constrained biophysical sequence optimizers

A Chen, SD Stanton, RG Alberstein, AM Watkins… - arxiv preprint arxiv …, 2024 - arxiv.org
Large language models (LLMs) have recently shown significant potential in various
biological tasks such as protein engineering and molecule design. These tasks typically …

Understanding the Logic of Direct Preference Alignment through Logic

K Richardson, V Srikumar, A Sabharwal - arxiv preprint arxiv:2412.17696, 2024 - arxiv.org
Recent direct preference alignment algorithms (DPA), such as DPO, have shown great
promise in aligning large language models to human preferences. While this has motivated …

Declarative characterizations of direct preference alignment algorithms

K Richardson, V Srikumar, A Sabharwal - NeurIPS 2024 Workshop on … - openreview.net
Recent direct preference alignment algorithms (DPA), such as DPO, have shown great
promise in aligning large language models to human preferences. While this has motivated …