🏢 Courant Institute

Sketchy Moment Matching (SkMM) is a fast and theoretically sound data selection method for deep learning finetuning. By controlling variance-bias tradeoffs in high dimensions, SkMM drastically reduces…

New surrogate loss functions for learning-to-defer achieve Bayes-consistency, realizable H-consistency, and H-consistency bounds simultaneously, resolving open questions and improving L2D performance.

Novel surrogate losses with label-independent H-consistency bounds enable stronger guarantees for multi-label learning.

A new formula unifies lazy and active neural network training regimes, revealing a mixed regime that combines their strengths for faster convergence and low-rank bias.

This paper reveals a universal square-root growth rate for H-consistency bounds of smooth surrogate losses in classification, significantly advancing our understanding of loss function selection.