Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling

TL;DR

This paper empirically compares various gated recurrent units like LSTM and GRU in sequence modeling tasks, demonstrating their superiority over traditional units and showing GRU's performance is comparable to LSTM.

Contribution

It provides an empirical evaluation of gated recurrent units, highlighting their effectiveness in sequence modeling and comparing LSTM and GRU performance.

Findings

Gated units outperform traditional tanh units.

GRU performs comparably to LSTM.

Advanced units improve sequence modeling results.

Abstract

In this paper we compare different types of recurrent units in recurrent neural networks (RNNs). Especially, we focus on more sophisticated units that implement a gating mechanism, such as a long short-term memory (LSTM) unit and a recently proposed gated recurrent unit (GRU). We evaluate these recurrent units on the tasks of polyphonic music modeling and speech signal modeling. Our experiments revealed that these advanced recurrent units are indeed better than more traditional recurrent units such as tanh units. Also, we found GRU to be comparable to LSTM.