Paying More Attention to Attention: Improving the Performance of Convolutional Neural Networks via Attention Transfer
论文地址:https://arxiv.org/abs/1612.03928
github地址:https://github.com/szagoruyko/attention-transfer
published: ICLR 2017
Abstract
student CNN 通过模仿teacher CNN的attention map 来提高性能
Introduction
One of popular hypothesis: There are non-attentional and attentional perception processes.
当进入一个未知场景后,首先是non-attention,Non-attentional processes help to observe a scene in general and gather high-level information。然后, control the attention processes and navigate to a certain part of the scene. This implies that different observers with different knowledge, different goals, and therefore different attentional strategies can literally see the same scene differently.
本文创新点
1)activation-based patial attention maps
这个方法能够对网络性能带来一定提升。
2)gradient-based patial attention maps
这个方法的效果不大