Strengthening Layer Interaction via Dynamic Layer Attention

TL;DR

This paper introduces Dynamic Layer Attention (DLA), a novel architecture that enhances layer interaction in neural networks by enabling dynamic context feature extraction, leading to improved performance in image recognition and object detection.

Contribution

The paper proposes a dynamic layer attention mechanism with a dual-path architecture and a new recurrent block, the Dynamic Sharing Unit, to improve layer interaction in neural networks.

Findings

DLA outperforms state-of-the-art methods in image recognition.

DLA improves object detection accuracy.

The DSU block is an efficient plugin for dynamic feature extraction.

Abstract

In recent years, employing layer attention to enhance interaction among hierarchical layers has proven to be a significant advancement in building network structures. In this paper, we delve into the distinction between layer attention and the general attention mechanism, noting that existing layer attention methods achieve layer interaction on fixed feature maps in a static manner. These static layer attention methods limit the ability for context feature extraction among layers. To restore the dynamic context representation capability of the attention mechanism, we propose a Dynamic Layer Attention (DLA) architecture. The DLA comprises dual paths, where the forward path utilizes an improved recurrent neural network block, named Dynamic Sharing Unit (DSU), for context feature extraction. The backward path updates features using these shared context representations. Finally, the attention mechanism is applied to these dynamically refreshed feature maps among layers. Experimental results demonstrate the effectiveness of the proposed DLA architecture, outperforming other state-of-the-art methods in image recognition and object detection tasks. Additionally, the DSU block has been evaluated as an efficient plugin in the proposed DLA architecture.The code is available at https://github.com/tunantu/Dynamic-Layer-Attention.