Layer-wise interpretation of deep neural networks using identity initialization

Shohei Kubota, Hideaki Hayashi, Tomohiro Hayase, Seiichi Uchida

研究成果: ジャーナルへの寄稿会議記事査読


The interpretability of neural networks (NNs) is a challenging but essential topic for transparency in the decision-making process using machine learning. One of the reasons for the lack of interpretability is random weight initialization, where the input is randomly embedded into a different feature space in each layer. In this paper, we propose an interpretation method for a deep multilayer perceptron, which is the most general architecture of NNs, based on identity initialization (namely, initialization using identity matrices). The proposed method allows us to analyze the contribution of each neuron to classification and class likelihood in each hidden layer. As a property of the identity-initialized perceptron, the weight matrices remain near the identity matrices even after learning. This property enables us to treat the change of features from the input to each hidden layer as the contribution to classification. Furthermore, we can separate the output of each hidden layer into a contribution map that depicts the contribution to classification and class likelihood, by adding extra dimensions to each layer according to the number of classes, thereby allowing the calculation of the recognition accuracy in each layer and thus revealing the roles of independent layers, such as feature extraction and classification.

ジャーナルICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
出版ステータス出版済み - 2021
イベント2021 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2021 - Virtual, Toronto, カナダ
継続期間: 6月 6 20216月 11 2021

!!!All Science Journal Classification (ASJC) codes

  • ソフトウェア
  • 信号処理
  • 電子工学および電気工学


「Layer-wise interpretation of deep neural networks using identity initialization」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。