Rethinking Batch Sample Relationships for Data Representation: A Batch-Graph Transformer Based Approach

Abstract

Exploring sample relationships within each mini-batch has shown great potential for learning image representations. Existing works generally adopt the regular Transformer to model the visual content relationships, ignoring the cues of semantic/label correlations between samples. Also, they generally adopt the ‘full’ self-attention mechanism which are obviously redundant and also sensitive to the noisy samples. To overcome these issues, in this paper, we design a simple yet flexible Batch-Graph Transformer (BGFormer) for mini-batch sample representations by deeply capturing the relationships of image samples from both visual and semantic perspectives. BGFormer has three main aspects. (1) It employs a flexible graph model, termed <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Batch Graph</i> to jointly encode both visual and semantic relationships of samples within each mini-batch. (2) It explores the neighborhood relationships of samples by borrowing the idea of sparse graph representation which thus performs robustly, w.r.t., noisy samples. (3) It devises a novel specific Transformer architecture that mainly adopts <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">dual</i> structure-constrained self-attention (SSA), together with graph normalization, FFN, etc, to carefully exploit the batch graph information for sample tokens (nodes) representations. As an application, we apply BGFormer to the metric learning tasks. Extensive experiments on four popular datasets demonstrate the effectiveness of the proposed model.