Methods for Interpreting and Understanding Deep Neural Networks
arXiv:1706.07979 · doi:10.1016/j.dsp.2017.10.011
Abstract
This paper provides an entry point to the problem of interpreting a deep neural network model and explaining its predictions. It is based on a tutorial given at ICASSP 2017. It introduces some recently proposed techniques of interpretation, along with theory, tricks and recommendations, to make most efficient use of these techniques on real data. It also discusses a number of practical applications.
14 pages, 10 figures
References in corpus (4)
Cited by in corpus (17)
- Explainable Artificial Intelligence: Understanding, Visualizing and Interpreting Deep Learning Models
- Massively Multilingual Neural Machine Translation in the Wild: Findings and Challenges
- Challenges of Real-World Reinforcement Learning
- Exploring Interpretable LSTM Neural Networks over Multi-Variable Data
- Concise Fuzzy System Modeling Integrating Soft Subspace Clustering and Sparse Learning
- Understanding and Comparing Deep Neural Networks for Age and Gender Classification
- A Rate-Distortion Framework for Explaining Neural Network Decisions
- Deep Representation Learning for Social Network Analysis
- The Convergence of Machine Learning and Communications
- Saliency-driven Word Alignment Interpretation for Neural Machine Translation
- Exploring text datasets by visualizing relevant words
- GAN-based Generation and Automatic Selection of Explanations for Neural Networks
- A Simple Saliency Method That Passes the Sanity Checks
- Discovering topics in text datasets by visualizing relevant words
- Teaching AI to Explain its Decisions Using Embeddings and Multi-Task Learning
- An AI-Augmented Lesion Detection Framework For Liver Metastases With Model Interpretability
- Regression Concept Vectors for Bidirectional Explanations in Histopathology