Yangfu Li

PhD student, East China Normal University

1 paper at NeurIPS 2025

Homepage· OpenReview· Semantic Scholar· Google Scholar

Poster Session 4

1 paper

Thursday, December 4, 2025 · 4:30 PM → 7:30 PM

Exhibit Hall C,D,E

Why 1 + 1 < 1 in Visual Token Pruning: Beyond Naive Integration via Multi-Objective Balanced Covering

#4718 · Yangfu Li, Hongjian Zhan, Tianyi Chen, Qi Liu, Yu-Jie Xiong, Yue Lu

This study reveals and quantifies key factors in visual token pruning from a geometric covering perspective, and proposes Multi-Objective Balanced Covering, which significantly accelerates MLLM reasoning with negligible performance loss.