arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2409.08712cs.LGcs.AIcs.CLcs.CV
收藏

神经网络中知识的逐层变化

Layerwise Change of Knowledge in Neural Networks

Shanghai Jiao Tong University(上海交通大学);Nanjing University of Science and Technology(南京理工大学);Zhejiang University(浙江大学);Stevens Institute of Technology(史蒂文斯理工学院)

Xu Cheng, Lei Cheng, Zhaoran Peng, Yang Xu, Tian Han, Quanshi Zhang

AI总结:

该研究拓展交互作用定义,首次提取并量化追踪DNN前向传播各层新出现与被遗忘的交互作用,为理解其学习行为、泛化能力变化及特征表示不稳定性提供新视角。

摘要:

本文旨在解释深度神经网络(DNN)在前向传播过程中如何通过各层逐步提取新知识并遗忘噪声特征。截至目前,尽管学界对DNN所编码知识的定义尚未达成共识,但已有研究通过一系列数学证据,将交互作用视为DNN编码的符号化基本推理模式。本文拓展了交互作用的定义,并首次提取了中间层所编码的交互作用。我们对前向传播过程中每一层新出现的交互作用和被遗忘的交互作用进行量化与追踪,为理解DNN的学习行为提供了新的视角。交互作用的逐层变化还揭示了DNN泛化能力的变化以及特征表示的不稳定性。

英文摘要:

This paper aims to explain how a deep neural network (DNN) gradually extracts new knowledge and forgets noisy features through layers in forward propagation. Up to now, although the definition of knowledge encoded by the DNN has not reached a consensus, Previous studies have derived a series of mathematical evidence to take interactions as symbolic primitive inference patterns encoded by a DNN. We extend the definition of interactions and, for the first time, extract interactions encoded by intermediate layers. We quantify and track the newly emerged interactions and the forgotten interactions in each layer during the forward propagation, which shed new light on the learning behavior of DNNs. The layer-wise change of interactions also reveals the change of the generalization capacity and instability of feature representations of a DNN.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑