Deep Reinforced Attention Learning for Quality-Aware Visual Recognition
This work addresses the challenge of fully exploiting attention mechanisms in visual recognition for researchers and practitioners, offering a universal method to boost performance, though it is incremental as it builds on existing attention modules.
The paper tackles the problem of improving attention modules in convolutional neural networks by introducing a reinforcement learning method that evaluates and optimizes attention map quality, resulting in enhanced recognition performance across various benchmarks.
In this paper, we build upon the weakly-supervised generation mechanism of intermediate attention maps in any convolutional neural networks and disclose the effectiveness of attention modules more straightforwardly to fully exploit their potential. Given an existing neural network equipped with arbitrary attention modules, we introduce a meta critic network to evaluate the quality of attention maps in the main network. Due to the discreteness of our designed reward, the proposed learning method is arranged in a reinforcement learning setting, where the attention actors and recurrent critics are alternately optimized to provide instant critique and revision for the temporary attention representation, hence coined as Deep REinforced Attention Learning (DREAL). It could be applied universally to network architectures with different types of attention modules and promotes their expressive ability by maximizing the relative gain of the final recognition performance arising from each individual attention module, as demonstrated by extensive experiments on both category and instance recognition benchmarks.