Yilun Zhu

Even state-of-the-art LMMs struggle with reasoning-informed visual editing, with the best model only achieving 56.6% accuracy on a new comprehensive benchmark.

Finite-sample performance guarantees for label-noise transition matrix estimation are now achievable by bypassing class-posterior estimation through one-sided selective classification.

Deductive stereotyping in LLMs leads to biased reasoning, but targeted injection phrases can dramatically enhance fairness across diverse tasks.

A single biased example can completely undermine the alignment of large language models, revealing a critical flaw in their post-training safety mechanisms.

Current video MLLMs struggle to grasp fleeting visual events, with top models barely surpassing 39% accuracy on critical momentary tasks.