NeuroView-RNN: It's About Time

TL;DR

NeuroView-RNN introduces a family of RNN architectures that enhance interpretability by quantifying the contribution of each time step to decisions, while also often improving accuracy across various time-series datasets.

Contribution

It presents a novel RNN design that makes the importance of each hidden state interpretable through a linear classifier, offering both explainability and improved performance.

Findings

NeuroView-RNN can quantify the importance of each time step.

It often achieves higher accuracy than standard RNNs.

Demonstrated effectiveness on diverse time-series datasets.

Abstract

Recurrent Neural Networks (RNNs) are important tools for processing sequential data such as time-series or video. Interpretability is defined as the ability to be understood by a person and is different from explainability, which is the ability to be explained in a mathematical formulation. A key interpretability issue with RNNs is that it is not clear how each hidden state per time step contributes to the decision-making process in a quantitative manner. We propose NeuroView-RNN as a family of new RNN architectures that explains how all the time steps are used for the decision-making process. Each member of the family is derived from a standard RNN architecture by concatenation of the hidden steps into a global linear classifier. The global linear classifier has all the hidden states as the input, so the weights of the classifier have a linear mapping to the hidden states. Hence, from the weights, NeuroView-RNN can quantify how important each time step is to a particular decision. As a bonus, NeuroView-RNN also offers higher accuracy in many cases compared to the RNNs and their variants. We showcase the benefits of NeuroView-RNN by evaluating on a multitude of diverse time-series datasets.