Recurrent Neural Networks (RNN) are widely used for learning sequences in applications such as EEG classification. Complex RNNs could be hardly deployed on wearable devices due to their computation and memory-intensive processing patterns. Generally, reduction in precision leads much more efficiency and binarized RNNs are introduced as energy-efficient solutions. However, naive binarization methods lead to significant accuracy loss in EEG classification. In this paper, we propose a multi-level binarized LSTM, which significantly reduces computations whereas ensuring an accuracy pretty close to the full precision LSTM. Our method reduces the delay of the 3-bit LSTM cell operation 47* with less than 0.01% accuracy loss.
@article{arxiv.2004.08914,
title = {MuBiNN: Multi-Level Binarized Recurrent Neural Network for EEG signal Classification},
author = {Seyed Ahmad Mirsalari and Sima Sinaei and Mostafa E. Salehi and Masoud Daneshtalab},
journal= {arXiv preprint arXiv:2004.08914},
year = {2020}
}
Comments
To appear in IEEE International Symposium on Circuits & Systems in 2020. arXiv admin note: text overlap with arXiv:1807.04093 by other authors