An always-on 3.8μJ/86% CIFAR-10 mixed-signal binary CNN processor with all memory on chip in 28nm CMOS
In: 2018 IEEE International Solid - State Circuits Conference - (ISSCC), 2018-02-01
Online
unknown
Zugriff:
The trend of pushing deep learning from cloud to edge due to concerns of latency, bandwidth, and privacy has created demand for low-energy deep convolutional neural networks (CNNs). The single-layer classifier in [1] achieves sub-nJ operation, but is limited to moderate accuracy on low-complexity tasks (90% on MNIST). Larger CNN chips provide dataflow computing for high-complexity tasks (AlexNet) at mJ energy [2], but edge deployment remains a challenge due to off-chip DRAM access energy. This paper describes a mixed-signal binary CNN processor that performs image classification of moderate complexity (86% on CIFAR-10) and employs near-memory computing to achieve a classification energy of 3.8μJ, a 40x improvement over TrueNorth [3]. We accomplish this using (1) the BinaryNet algorithm for CNNs with weights and activations constrained to +1/−1 [4], which drastically simplifies multiplications (XNOR) and allows integrating all memory on-chip; (2) an energy-efficient switched-capacitor (SC) neuron that addresses BinaryNet's challenge of wide vector summation; (3) architectural parallelism, parameter reuse, and locality.
Titel: |
An always-on 3.8μJ/86% CIFAR-10 mixed-signal binary CNN processor with all memory on chip in 28nm CMOS
|
---|---|
Autor/in / Beteiligte Person: | Verhelst, Marian ; Bankman, Daniel ; Murmann, Boris ; Yang, Lita ; Moons, Bert |
Link: | |
Zeitschrift: | 2018 IEEE International Solid - State Circuits Conference - (ISSCC), 2018-02-01 |
Veröffentlichung: | IEEE, 2018 |
Medientyp: | unknown |
DOI: | 10.1109/isscc.2018.8310264 |
Schlagwort: |
|
Sonstiges: |
|