MindReading: An Ultra-Low-Power Photonic Accelerator for EEG-based Human Intention Recognition
MindReading: An Ultra-Low-Power Photonic Accelerator for EEG-based Human Intention Recognition
复制标题
DOI：
10.1109/asp-dac47756.2020.9045333
复制
发表时间：
2020-01
期刊：
影响因子：
--
通讯作者：
Qian Lou;Wenyang Liu;Weichen Liu;Feng Guo;Lei Jiang
中科院分区：
文献类型：
--
作者：
Qian Lou;Wenyang Liu;Weichen Liu;Feng Guo;Lei Jiang
一种基于脑电图（EEG）的脑机接口（BCI）系统可以极大地改善运动障碍患者的生活质量。构建由多个卷积层、LSTM层和全连接层组成的深度神经网络，对脑电信号进行解码，最大限度地提高人类意图识别的准确率。然而，现有的FPGA、ASIC、ReRAM和光子加速器在处理实时意图识别时无法保持足够的电池寿命。在本文中，我们提出了一种超低功率光子加速器，MindReading，用于人类意图识别，仅通过低位宽加法和移位操作。与之前的神经网络加速器相比，为了保持实时处理吞吐量，MindReading将功耗降低了62.7%，每瓦特吞吐量提高了168%。
A scalp-recording electroencephalography (EEG)-based brain-computer interface (BCI) system can greatly improve the quality of life for people who suffer from motor disabilities. Deep neural networks consisting of multiple convolutional, LSTM and fully-connected layers are created to decode EEG signals to maximize the human intention recognition accuracy. However, prior FPGA, ASIC, ReRAM and photonic accelerators cannot maintain sufficient battery lifetime when processing realtime intention recognition. In this paper, we propose an ultra-low-power photonic accelerator, MindReading, for human intention recognition by only low bit-width addition and shift operations. Compared to prior neural network accelerators, to maintain the real-time processing throughput, MindReading reduces the power consumption by 62.7% and improves the throughput per Watt by 168%.