Di Wu

UMI-Bench 1.0 reveals that standardized, reproducible evaluations can significantly improve our understanding of UMI-style manipulation policies in real-world settings.

Achieving high-dimensional audio representations without sacrificing generative quality, the F3-Tokenizer redefines the boundaries of audio autoencoding.

AI reviews can identify significant issues in paper drafts that human reviewers might overlook, raising important questions about the future of academic peer review.