Quaternion Recurrent Neural Networks

TL;DR

This paper introduces quaternion-based recurrent neural networks (QRNN and QLSTM) that better model multi-dimensional sequential data, achieving improved performance and parameter efficiency in speech recognition tasks.

Contribution

The paper presents novel quaternion RNN and LSTM architectures that incorporate internal structural dependencies using quaternion algebra, enhancing modeling capabilities and efficiency.

Findings

QRNN and QLSTM outperform traditional RNN and LSTM in speech recognition.

They reduce the number of parameters by up to 3.3 times.

Achieve better accuracy with more compact models.

Abstract

Recurrent neural networks (RNNs) are powerful architectures to model sequential data, due to their capability to learn short and long-term dependencies between the basic elements of a sequence. Nonetheless, popular tasks such as speech or images recognition, involve multi-dimensional input features that are characterized by strong internal dependencies between the dimensions of the input vector. We propose a novel quaternion recurrent neural network (QRNN), alongside with a quaternion long-short term memory neural network (QLSTM), that take into account both the external relations and these internal structural dependencies with the quaternion algebra. Similarly to capsules, quaternions allow the QRNN to code internal dependencies by composing and processing multidimensional features as single entities, while the recurrent operation reveals correlations between the elements composing the sequence. We show that both QRNN and QLSTM achieve better performances than RNN and LSTM in a realistic application of automatic speech recognition. Finally, we show that QRNN and QLSTM reduce by a maximum factor of 3.3x the number of free parameters needed, compared to real-valued RNNs and LSTMs to reach better results, leading to a more compact representation of the relevant information.