Title | ||
---|---|---|
A Multilayer-Learning Current-Mode Neuromorphic System with Analog-Error Compensation. |
Abstract | ||
---|---|---|
Internet-of-things (IoT) applications that use machine-learning algorithms have increased the demand for application-specific energy-efficient hardware that can perform both learning and inference tasks to adapt to endpoint users or environmental changes. This paper presents a multilayer-learning neuromorphic system with analog-based multiplier-accumulator (MAC), which can learn training data by stochastic gradient descent algorithm. As a component of the proposed system, a current-mode MAC processor, fabricated in 28-nm CMOS technology, performs both forward and backward processing in a crossbar structure of 500 × 500 6-bit transposable SRAM arrays. The proposed system is verified in a two-layer neural network by using two prototype chips and an FPGA. Without any calibration circuit for the analog-based MAC, the proposed system compensates for non-idealities from analog operations by learning training data with the analog-based MAC. With 1-bit (+1, 0, -1) batch update of 6-bit synaptic weights, the proposed system achieves a recognition rate of 96.6 % with a peak energy efficiency of 2.99 TOPS/W (1 OP = one unsigned 8-bit × signed 6-bit MAC operation) in the classification of the MNIST dataset. |
Year | DOI | Venue |
---|---|---|
2019 | 10.1109/TBCAS.2019.2929696 | IEEE transactions on biomedical circuits and systems |
Keywords | Field | DocType |
Hardware,Neuromorphics,Pulse width modulation,Neural networks,Energy efficiency,Artificial intelligence,Synapses | Stochastic gradient descent,MNIST database,Computer science,Field-programmable gate array,Neuromorphic engineering,Static random-access memory,Electronic engineering,CMOS,Artificial neural network,Computer hardware,Crossbar switch | Journal |
Volume | Issue | ISSN |
13 | 5 | 1932-4545 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
7 |
Name | Order | Citations | PageRank |
---|---|---|---|
Hyunwoo Son | 1 | 11 | 4.43 |
Hwasuk Cho | 2 | 2 | 3.12 |
Jungho Lee | 3 | 22 | 4.55 |
Seongun Bae | 4 | 0 | 0.34 |
Byungsub Kim | 5 | 165 | 37.71 |
Hong-june Park | 6 | 465 | 72.93 |
Jae-yoon Sim | 7 | 508 | 83.58 |