SOURCE-LINKED INTELLIGENCE

Who Speaks for the Pruned? Visual Token Pruning as Coverage Optimization

arXiv · AI, language, vision and robotics · article · Sep 2, 2026 · UTC

Visual token pruning reduces the inference cost of vision-language models (VLMs), but most methods only ask which tokens to keep. This retained-token view can keep redundant high-scoring tokens while leaving discarded evidence without a close representative. We propose CoverPruner, a training-free pruner that asks the complementary demand-side question: after a token is removed, which surviving original token represents it for the target VLM? CoverPruner formulates pruning as Representational Coverage Maximization (RCM), covering the full projected visual-token set with query-weighted demand.

Read original source ↗ Open in workspace

recordType

paper

region

Global