paper-with-me

Papers

Rényi Attention Entropy for Patch Pruning

2026-04-04 · Hiroaki Aizawa, Yuki Igaue arxiv

Transformers are strong baselines in both vision and language because self-attention captures long-range dependencies across tokens. However, the cost of self-attention grows quadratically with the number of tokens. Patch pruning mitigates this cost by estimating per-patch importance and removing redundant patches. To identify informative patches for pruning, we introduce a criterion based on the Shannon entropy of the attention distribution. Low-entropy patches, which receive selective and concentrated attention, are kept as important, while high-entropy patches with attention spread across many locations are treated as redundant. We also extend the criterion from Shannon to Rényi entropy, which emphasizes sharp attention peaks and supports pruning strategies that adapt to task needs and computational limits. In experiments on fine-grained image recognition, where patch selection is critical, our method reduced computation while preserving accuracy. Moreover, adjusting the pruning policy through the Rényi entropy measure yields further gains and improves the trade-off between accuracy and computation.

📄 PDF Abstract BibTeX arXiv:2604.03803

Code (0)

등록된 구현이 없습니다.

Tasks

Fine-Grained Image Recognition

Similar Papers 제목 키워드 기반

Rényi Entropy: A New Token Pruning Metric for Vision Transformers

2026-03-29 · Wei-Yuan Su, Ruijie Zhang, Zheng Zhang arxiv

Vision Transformers (ViTs) achieve state-of-the-art performance but suffer from the $O(N^2)$ complexity of self-attention, making inference costly for high-resolution inputs. To address this bottleneck, token pruning has…

Evading Visual Aphasia: Contrastive Adaptive Semantic Token Pruning for Vision-Language Models

2026-05-10 · Jie Ma, Yihang Liu, Zhike Qiu, Jiayi Ji 외 arxiv

Are low-attention visual tokens truly redundant in vision-language reasoning? Existing pruning methods often assume so, ranking visual tokens by shallow text-to-image attention and discarding low-scoring patches to accel…

Visual Grounding

Dispatch-Aware Ragged Attention for Pruned Vision Transformers

2026-04-16 · Seifeldin Abdellatif, Ahmad Almasri arxiv

Token pruning methods for Vision Transformers (ViTs) promise quadratic reductions in attention FLOPs by dropping uninformative patches. Yet standard variable-length attention APIs -- including FlashAttention-2's varlen a…

APP: Accelerated Path Patching with Task-Specific Pruning

2025-11-07 · Frauke Andersen, William Rudman, Ruochen Zhang, Carsten Eickhoff arxiv

Circuit discovery is a key step in many mechanistic interpretability pipelines. Current methods, such as Path Patching, are computationally expensive and have limited in-depth circuit analysis for smaller models. In this…

Patch Pruning Strategy Based on Robust Statistical Measures of Attention Weight Diversity in Vision Transformers

2025-07-25 · Yuki Igaue, Hiroaki Aizawa arxiv

Multi-head self-attention is a distinctive feature extraction mechanism of vision transformers that computes pairwise relationships among all input patches, contributing significantly to their high performance. However, …

Computational Efficiency