Input Switched Affine Networks: An RNN Architecture Designed for Interpretability

11/28/2016
by   Jakob N. Foerster, et al.
0

There exist many problem domains where the interpretability of neural network models is essential for deployment. Here we introduce a recurrent architecture composed of input-switched affine transformations - in other words an RNN without any explicit nonlinearities, but with input-dependent recurrent weights. This simple form allows the RNN to be analyzed via straightforward linear methods: we can exactly characterize the linear contribution of each input to the model predictions; we can use a change-of-basis to disentangle input, output, and computational hidden unit subspaces; we can fully reverse-engineer the architecture's solution to a simple task. Despite this ease of interpretation, the input switched affine network achieves reasonable performance on a text modeling tasks, and allows greater computational efficiency than networks with standard nonlinearities.

READ FULL TEXT

page 4

page 5

page 6

page 8

11/06/2017

Neural Speed Reading via Skim-RNN

Inspired by the principles of speed reading, we introduce Skim-RNN, a re...
09/26/2017

Input-to-Output Gate to Improve RNN Language Models

This paper proposes a reinforcing method that refines the output layers ...
03/03/2016

Training Input-Output Recurrent Neural Networks through Spectral Methods

We consider the problem of training input-output recurrent neural networ...
06/09/2019

Attention-based Conditioning Methods for External Knowledge Integration

In this paper, we present a novel approach for incorporating external kn...
06/02/2021

Framing RNN as a kernel method: A neural ODE approach

Building on the interpretation of a recurrent neural network (RNN) as a ...
01/25/2016

Survey on the attention based RNN model and its applications in computer vision

The recurrent neural networks (RNN) can be used to solve the sequence to...
03/27/2019

On Attribution of Recurrent Neural Network Predictions via Additive Decomposition

RNN models have achieved the state-of-the-art performance in a wide rang...