Comparing Fixed and Adaptive Computation Time for Recurrent Neural Networks
Neural and Evolutionary Computing
2018-03-23 v1 Machine Learning
Abstract
Adaptive Computation Time for Recurrent Neural Networks (ACT) is one of the most promising architectures for variable computation. ACT adapts to the input sequence by being able to look at each sample more than once, and learn how many times it should do it. In this paper, we compare ACT to Repeat-RNN, a novel architecture based on repeating each sample a fixed number of times. We found surprising results, where Repeat-RNN performs as good as ACT in the selected tasks. Source code in TensorFlow and PyTorch is publicly available at https://imatge-upc.github.io/danifojo-2018-repeatrnn/
Keywords
Cite
@article{arxiv.1803.08165,
title = {Comparing Fixed and Adaptive Computation Time for Recurrent Neural Networks},
author = {Daniel Fojo and Víctor Campos and Xavier Giro-i-Nieto},
journal= {arXiv preprint arXiv:1803.08165},
year = {2018}
}
Comments
Accepted as workshop paper at ICLR 2018