David Wipf

Researchers develop methods to better align language models with human preferences.

A new framework improves learning from structured, noisy data using diffusion principles.

SGFormer simplifies graph learning for efficiency and scalability.