Name
Abstract | ||
|---|---|---|
Deep neural network (DNN) accelerators received considerable attention in past years due to saved energy compared to mainstream hardware. Low-voltage operation of DNN accelerators allows to further reduce energy consumption significantly, however, causes bit-level failures in the memory storing the quantized DNN weights. In this paper, we show that a combination of robust fixed-point quantization, weight clipping, and random bit error training (RandBET) improves robustness against random bit errors in (quantized) DNN weights significantly. This leads to high energy savings from both low-voltage operation as well as low-precision quantization. Our approach generalizes across operating voltages and accelerators, as demonstrated on bit errors from profiled SRAM arrays. We also discuss why weight clipping alone is already a quite effective way to achieve robustness against bit errors. Moreover, we specifically discuss the involved trade-offs regarding accuracy, robustness and precision: Without losing more than 1% in accuracy compared to a normally trained 8-bit DNN, we can reduce energy consumption on CIFAR-10 by 20%. Higher energy savings of, e.g., 30%, are possible at the cost of 2.5% accuracy, even for 4-bit DNNs. |
Year | Venue | Keywords |
|---|---|---|
2021 | Conference on Machine Learning and Systems (MLSys) | Robustness (computer science),Bit error rate,Energy consumption,Quantization (signal processing),Efficient energy use,Static random-access memory,Artificial neural network,Computer engineering,Quantization (physics),Computer science |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
0 | 4 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| David Stutz | 1 | 0 | 0.34 |
| Nandhini Chandramoorthy | 2 | 0 | 0.34 |
| Matthias Hein | 3 | 663 | 62.80 |
| Bernt Schiele | 4 | 12901 | 971.29 |