Stochastic computing (SC) has been used for lowcost and low power implementation of neural networks. Inherent inaccuracy and long latency of processing random bit-streams have made prior SC-based implementations inefficient compared to conventional fixed-point designs. Random or pseudo-random bitstreams often need to be processed for a very long time to produce acceptable results. This long latency leads to a significantly higher energy consumption than binary design counterparts. Low-discrepancy sequences have been recently used for fast-converging deterministic computation with stochastic constructs. In this work, we propose a low-cost, low-latency, and energy-efficient implementation of convolutional neural networks based on low-discrepancy deterministic bit-streams. Experimental results show a significant reduction in the energy consumption compared to previous random bitstream-based implementations and to the optimized fixed-point design with no quality degradation.
|Original language||English (US)|
|Title of host publication||Proceedings of the 2019 Design, Automation and Test in Europe Conference and Exhibition, DATE 2019|
|Publisher||Institute of Electrical and Electronics Engineers Inc.|
|Number of pages||6|
|State||Published - May 14 2019|
|Event||22nd Design, Automation and Test in Europe Conference and Exhibition, DATE 2019 - Florence, Italy|
Duration: Mar 25 2019 → Mar 29 2019
|Name||Proceedings of the 2019 Design, Automation and Test in Europe Conference and Exhibition, DATE 2019|
|Conference||22nd Design, Automation and Test in Europe Conference and Exhibition, DATE 2019|
|Period||3/25/19 → 3/29/19|
Bibliographical noteFunding Information:
ACKNOWLEDGMENT This work was supported in part by National Science Foundation grant no. CCF-1438286. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the NSF.
This work was supported in part by National Science Foundation grant no. CCF-1438286.
© 2019 EDAA.
- Convolutional neural networks
- bitstream processing
- energy-efficient design
- low-cost design
- stochastic computing