Explaining Deep Neural Networks using Unsupervised Clustering
arXiv:2007.07477
Abstract
We propose a novel method to explain trained deep neural networks (DNNs), by distilling them into surrogate models using unsupervised clustering. Our method can be applied flexibly to any subset of layers of a DNN architecture and can incorporate low-level and high-level information. On image datasets given pre-trained DNNs, we demonstrate the strength of our method in finding similar training samples, and shedding light on the concepts the DNNs base their decisions on. Via user studies, we show that our model can improve the user trust in model's prediction.
References in corpus (7)
- Learning Important Features Through Propagating Activation Differences
- Deep Clustering for Unsupervised Learning of Visual Features
- Distilling a Neural Network Into a Soft Decision Tree
- Unsupervised Deep Embedding for Clustering Analysis
- On Completeness-aware Concept-Based Explanations in Deep Neural Networks
- Stop Explaining Black Box Machine Learning Models for High Stakes Decisions and Use Interpretable Models Instead
- ProtoAttend: Attention-Based Prototypical Learning