paper-with-me

Papers

Inducing Spatial Locality in Vision Transformers through the Training Protocol

2026-05-11 · Eduardo Santiago Toledo, Asael Fabian Martínez arxiv

We investigate whether the training protocol can induce spatial locality in the early layers of a Vision Transformer (ViT) trained from scratch, without large-scale pretraining. Keeping the architecture and optimization procedure fixed, we compare a Baseline protocol with a Modern protocol (AutoAugment/ColorJitter, CutMix, and Label Smoothing) on CIFAR-10, CIFAR-100, and Tiny-ImageNet, characterizing each attention head via Mean Attention Distance (MAD) and normalized entropy. Across all three datasets, the Modern protocol produces more local and more concentrated attention in early layers; on CIFAR-100, the minimum MAD drops from 0.316 (Baseline) to 0.008 (Modern). To identify the source of this effect, we conduct an ablation study on CIFAR-100 by adding or removing each component individually. The results identify CutMix as the determining component within our experiments: all conditions with CutMix exhibit MAD 0.024, while all conditions without CutMix remain at MAD 0.210. AutoAugment and Label Smoothing show no independent effect on locality. Taken together, these findings suggest that the pressure to classify from partial image regions, induced by CutMix, can promote the emergence of local attention in Vision Transformers.

📄 PDF Abstract BibTeX arXiv:2605.16390

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

MaiT: integrating spatial locality into image transformers with attention masks

2021-09-29 · Ling Li, Ali Shafiee, Joseph H Hassoun

Though image transformers have shown competitive results with convolutional neural networks in computer vision tasks, lacking inductive biases such as locality still poses problems in terms of model efficiency especially…

MaiT: Leverage Attention Masks for More Efficient Image Transformers

2022-07-06 · Ling Li, Ali Shafiee Ardestani, Joseph Hassoun

Though image transformers have shown competitive results with convolutional neural networks in computer vision tasks, lacking inductive biases such as locality still poses problems in terms of model efficiency especially…

Grid-Preserving Knowledge Distillation: Transferring Convolutional Inductive Bias to Vision Transformers under Data Scarcity

2026-08-11 · Junyong Choi, Cheolhyeon Park, Jaehoon Cho arxiv

Vision Transformers demonstrate remarkable global modeling capacity but often underperform in data-scarce regimes. Distilling convolutional inductive biases from a CNN teacher provides an effective remedy while leaving t…

Knowledge DistillationGeneral Knowledge

LocalViT: Bringing Locality to Vision Transformers

2021-04-12 · Yawei Li, Kai Zhang, JieZhang Cao, Radu Timofte 외

We study how to introduce locality mechanisms into vision transformers. The transformer network originates from machine translation and is particularly good at modelling long-range dependencies within a long sequence. Al…

Image Classification

Locality Alignment Improves Vision-Language Models

2024-10-14 · Ian Covert, Tony Sun, James Zou, Tatsunori Hashimoto

Vision language models (VLMs) have seen growing adoption in recent years, but many still struggle with basic spatial reasoning errors. We hypothesize that this is due to VLMs adopting pre-trained vision backbones, specif…

Semantic SegmentationSpatial Reasoning