Paper
Quaternion Recurrent Neural Networks
Recurrent neural networks (RNNs) are powerful architectures to model\nsequential data, due to their capability to learn short and long-term\ndependencies between the basic elements of a sequence. Nonetheless, popular\ntasks such as speech or images recognition, involve multi-dimensional input\nfeatures that are characterized by strong internal dependencies between the\ndimensions of the input vector. We propose a novel quaternion recurrent neural\nnetwork (QRNN), alongside with a quaternion long-short term memory neural\nnetwork (QLSTM), that take into account both the external relations and these\ninternal structural dependencies with the quaternion algebra. Similarly to\ncapsules, quaternions allow the QRNN to code internal dependencies by composing\nand processing multidimensional features as single entities, while the\nrecurrent operation reveals correlations between the elements composing the\nsequence. We show that both QRNN and QLSTM achieve better performances than RNN\nand LSTM in a realistic application of automatic speech recognition. Finally,\nwe show that QRNN and QLSTM reduce by a maximum factor of 3.3x the number of\nfree parameters needed, compared to real-valued RNNs and LSTMs to reach better\nresults, leading to a more compact representation of the relevant information.\n
Authors: Parcollet, Titouan · Ravanelli, Mirco · Morchid, Mohamed · Linarès, Georges · Trabelsi, Chiheb · De Mori, Renato · Bengio, Yoshua