Shapley Interpretation and Activation in Neural Networks
arXiv:1909.06143
Abstract
We propose a novel Shapley value approach to help address neural networks' interpretability and "vanishing gradient" problems. Our method is based on an accurate analytical approximation to the Shapley value of a neuron with ReLU activation. This analytical approximation admits a linear propagation of relevance across neural network layers, resulting in a simple, fast and sensible interpretation of neural networks' decision making process. We then derived a globally continuous and non-vanishing Shapley gradient, which can replace the conventional gradient in training neural network layers with ReLU activation, and leading to better training performance. We further derived a Shapley Activation (SA) function, which is a close approximation to ReLU but features the Shapley gradient. The SA is easy to implement in existing machine learning frameworks. Numerical tests show that SA consistently outperforms ReLU in training convergence, accuracy and stability.
arXiv comment11 pages, 7 figures
References in corpus (11)
- Adam: A Method for Stochastic Optimization
- On the difficulty of training Recurrent Neural Networks
- Methods for Interpreting and Understanding Deep Neural Networks
- Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs)
- Explainable Artificial Intelligence: Understanding, Visualizing and Interpreting Deep Learning Models
- How to Explain Individual Classification Decisions
- Self-Normalizing Neural Networks
- Learning how to explain neural networks: PatternNet and PatternAttribution
- Overcoming the vanishing gradient problem in plain recurrent networks
- Explaining Deep Neural Networks with a Polynomial Time Algorithm for Shapley Values Approximation
- Reduced Form Capital Optimization