START: A Generalized State Space Model with Saliency-Driven Token-Aware Transformation

TL;DR

START introduces a saliency-driven token-aware transformation within state space models, enabling better domain generalization by suppressing domain-specific features, achieving state-of-the-art results efficiently.

Contribution

The paper proposes a novel SSM-based architecture with saliency-based token-aware transformation (START) that improves domain generalization and outperforms existing methods.

Findings

START achieves SOTA performance on five benchmarks.

START reduces domain discrepancy by suppressing domain-specific features.

START maintains linear complexity, ensuring computational efficiency.

Abstract

Domain Generalization (DG) aims to enable models to generalize to unseen target domains by learning from multiple source domains. Existing DG methods primarily rely on convolutional neural networks (CNNs), which inherently learn texture biases due to their limited receptive fields, making them prone to overfitting source domains. While some works have introduced transformer-based methods (ViTs) for DG to leverage the global receptive field, these methods incur high computational costs due to the quadratic complexity of self-attention. Recently, advanced state space models (SSMs), represented by Mamba, have shown promising results in supervised learning tasks by achieving linear complexity in sequence length during training and fast RNN-like computation during inference. Inspired by this, we investigate the generalization ability of the Mamba model under domain shifts and find that input-dependent matrices within SSMs could accumulate and amplify domain-specific features, thus hindering model generalization. To address this issue, we propose a novel SSM-based architecture with saliency-based token-aware transformation (namely START), which achieves state-of-the-art (SOTA) performances and offers a competitive alternative to CNNs and ViTs. Our START can selectively perturb and suppress domain-specific features in salient tokens within the input-dependent matrices of SSMs, thus effectively reducing the discrepancy between different domains. Extensive experiments on five benchmarks demonstrate that START outperforms existing SOTA DG methods with efficient linear complexity. Our code is available at https://github.com/lingeringlight/START.