Most feedforward convolutional neural networks spend roughly the same efforts for each pixel. Yet human visual recognition is an interaction between eye movements and spatial attention, which we will have several glimpses of an object in different regions. Inspired by this observation, we propose an end-to-end trainable Multi-Glimpse Network (MGNet) which aims to tackle the challenges of high computation and the lack of robustness based on recurrent downsampled attention mechanism. Specifically, MGNet sequentially selects task-relevant regions of an image to focus on and then adaptively combines all collected information for the final prediction. MGNet expresses strong resistance against adversarial attacks and common corruptions with less computation. Also, MGNet is inherently more interpretable as it explicitly informs us where it focuses during each iteration. Our experiments on ImageNet100 demonstrate the potential of recurrent downsampled attention mechanisms to improve a single feedforward manner. For example, MGNet improves 4.76% accuracy on average in common corruptions with only 36.9% computational cost. Moreover, while the baseline incurs an accuracy drop to 7.6%, MGNet manages to maintain 44.2% accuracy in the same PGD attack strength with ResNet-50 backbone. Our code is available at https://github.com/siahuat0727/MGNet.
翻译:多数进化变异神经网络花在每一个像素上的时间大致相同。 然而人类视觉识别是眼运动和空间关注之间的一种互动关系,我们将对不同区域的一个物体有几眼观察。 在这种观察的启发下,我们提议一个端到端可训练的多石墨网络(MGNet),目的是应对高计算的挑战和基于反复下降的吸引机制缺乏强力的问题。具体地说,MGNet按顺序选择一个图像相关任务区域,以关注并随后适应性地将所有收集到的最终预测信息结合起来。MGNet表示强烈抵制对抗对抗性攻击和常见腐败,而计算较少。此外,MGNet在每次循环期间明确告诉我们其重点位置时,因此具有内在的更可解释性。我们在图像网络100上进行的实验显示了经常性下降关注机制的潜力,以改进单一的反馈方式。例如,MGNet按普通腐败的平均4.76%的准确度选择,而只有36.9%的计算成本。此外,尽管基线显示攻击的准确性下降至7.6%,MGMGNet在MG/Resmax Riumal2 管理着MAC44中,MCD-com 。