Yixiong Zou

STD is proposed, a hierarchical token pruning framework that adapts token selection mechanisms to the functional role of each network stage, and introduces a Stability-Adaptive Trigger in deep layers to execute pruning only during semantically stable phases.

Low-quality modalities can quietly sabotage multi-modal networks because they receive minimal training gradients yet dominate test-time skip connections—a failure mode solved by explicitly filtering and rebalancing degraded feature streams.

Bridging the gap between domains with tailored noise and selective inpainting can revolutionize few-shot object detection performance.