Reliability Evaluation of Compressed Deep Learning Models
- 1 February 2020
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-5
Abstract
Neural networks are becoming deeper and more complex, making it harder to store and process such applications on systems with limited resources. Model pruning and data quantization are two effective ways to simplify the necessary hardware by compressing the network with relevant-only nodes and reducing the required data precision. Such optimizations, however, might come at a cost of reliability since critical nodes are now more exposed to faults and the network is more sensitive to small changes. In this work, we present an extensive empirical investigation of transient faults on compressed deep convolutional neural networks (CNNs). We evaluate the impact of a single bit flip over three CNN models with different sparsity configurations and integer-only quantizations. We show that pruning can increase the resilience of the system by 9x when compared to the dense model. Quantization can outperform the 32-bit floating-point baseline by adding 27.4x more resilience to the overall network and up to 108.7x when combined with pruning. This makes model compression an effective way to provide resilience to deep learning workloads during inference, mitigating the need for explicit error correction hardware.Keywords
This publication has 10 references indexed in Scilit:
- On the Resilience of RTL NN Accelerators: Fault Characterization and Mitigation2018
- Ares: A framework for quantifying the resilience of deep neural networks2018
- Understanding error propagation in deep learning neural network (DNN) accelerators and applications2017
- Designing Energy-Efficient Convolutional Neural Networks Using Energy-Aware Pruning2017
- Minerva: Enabling Low-Power, Highly-Accurate Deep Neural Network Accelerators2016
- Deep Residual Learning for Image Recognition2016
- Distillation as a Defense to Adversarial Perturbations Against Deep Neural Networks2016
- ImageNet Large Scale Visual Recognition ChallengeInternational Journal of Computer Vision, 2015
- Designing Reliable Systems from Unreliable Components: The Challenges of Transistor Variability and DegradationIEEE Micro, 2005
- Analysis of Fault Tolerance in Artificial Neural NetworksJournal of Parallel and Distributed Computing, 2001