Combating Textual Noise and Redundancy: Entropy-Aware Dense Visual Token Pruning
Researchers propose Entropy-Aware Dense Pruning (EADP) to improve visual token pruning in VLMs by addressing textual noise and feature fragmentation. EADP reformulates pruning as a structured compression problem.
Visual token pruning is a crucial strategy for accelerating VLMs by compressing redundant image patches, yet existing methods often fail to preserve critical cues under dense instructions and fine-grained queries. In this paper, we investigate this failure and identify two underlying bottlenecks: the widespread dispersion of textual noise that corrupts dense cross-modal scoring, and the feature fragmentation inherent to standard token selection. To address these issues, we propose Entropy-Aware Dense Pruning (EADP), a framework that reformulates pruning as a structured compression problem. EAD
Source: Combating Textual Noise and Redundancy: Entropy-Aware Dense Visual Token Pruning. Read the full piece at the source.
Summary and analysis generated by AI (groq). Always verify against the original sources.
Measuring the Economic Effects of AI - Economic Innovation Group
