Interpretable Embedding Procedure Knowledge Transfer via Stacked Principal Component Analysis and Graph Neural Network

TL;DR

This paper introduces an interpretable knowledge distillation method that uses PCA and graph neural networks to improve lightweight neural networks and visualize the embedding process.

Contribution

It proposes a novel interpretable embedding knowledge transfer method combining PCA and message passing neural networks for better distillation.

Findings

Improves CIFAR100 accuracy by 2.28% over SOTA

Provides visual interpretability of the embedding procedure

Demonstrates effectiveness of the proposed KD method

Abstract

Knowledge distillation (KD) is one of the most useful techniques for light-weight neural networks. Although neural networks have a clear purpose of embedding datasets into the low-dimensional space, the existing knowledge was quite far from this purpose and provided only limited information. We argue that good knowledge should be able to interpret the embedding procedure. This paper proposes a method of generating interpretable embedding procedure (IEP) knowledge based on principal component analysis, and distilling it based on a message passing neural network. Experimental results show that the student network trained by the proposed KD method improves 2.28% in the CIFAR100 dataset, which is higher performance than the state-of-the-art (SOTA) method. We also demonstrate that the embedding procedure knowledge is interpretable via visualization of the proposed KD process. The implemented code is available at https://github.com/sseung0703/IEPKT.