Dual Rectified Linear Units (DReLUs): A Replacement for Tanh Activation Functions in Quasi-Recurrent Neural Networks

by   Fréderic Godin, et al.

In this paper, we introduce a novel type of Rectified Linear Unit (ReLU), called a Dual Rectified Linear Unit (DReLU). A DReLU, which comes with an unbounded positive and negative image, can be used as a drop-in replacement for a tanh activation function in the recurrent step of Quasi-Recurrent Neural Networks (QRNNs) (Bradbury et al. (2017)). Similar to ReLUs, DReLUs are less prone to the vanishing gradient problem, they are noise robust, and they induce sparse activations. We independently reproduce the QRNN experiments of Bradbury et al. (2017) and compare our DReLU-based QRNNs with the original tanh-based QRNNs and Long Short-Term Memory networks (LSTMs) on sentiment classification and word-level language modeling. Additionally, we evaluate on character-level language modeling, showing that we are able to stack up to eight QRNN layers with DReLUs, thus making it possible to improve the current state-of-the-art in character-level language modeling over shallow architectures based on LSTMs.


page 1

page 2

page 3

page 4


Multiplicative Models for Recurrent Language Modeling

Recently, there has been interest in multiplicative recurrent neural net...

Pyramidal Recurrent Unit for Language Modeling

LSTMs are powerful tools for modeling contextual information, as evidenc...

Improving Language Modeling using Densely Connected Recurrent Neural Networks

In this paper, we introduce the novel concept of densely connected layer...

Quasi-Recurrent Neural Networks

Recurrent neural networks are a powerful tool for modeling sequential da...

Recurrent Highway Networks

Many sequential processing tasks require complex nonlinear transition fu...

Exploring the Naturalness of Buggy Code with Recurrent Neural Networks

Statistical language models are powerful tools which have been used for ...

Trellis Networks for Sequence Modeling

We present trellis networks, a new architecture for sequence modeling. O...