Reliability Analysis of a Spiking Neural Network Hardware Accelerator

Abstract

Despite the parallelism and sparsity in neural network models, their transfer into hardware unavoidably makes them susceptible to hardware-level faults. Hardware-level faults can occur either during manufacturing, such as physical defects and process-induced variations, or in the field due to environmental factors and aging. The performance under fault scenarios needs to be assessed so as to develop cost-effective fault-tolerance schemes. In this work, we assess the resilience characteristics of a hardware accelerator for Spiking Neural Networks (SNNs) designed in VHDL and implemented on an FPGA. The fault injection experiments pinpoint the parts of the design that need to be protected against faults, as well as the parts that are inherently fault-tolerant.

Compact Functional Testing for Neuromorphic Computing Circuits

“…Specifically now for SNNs, behavioral-level fault modeling is discussed in [33], [34]. Fault injection experiments at behavioral-level are described in [22], [23], [33], [35], [36], while fault injection experiments onto actual neuromorphic hardware are described in [16]. In [22], [23], symptom detectors are designed for the two main catastrophic fault mechanisms in SNNs.…”

“…The common conclusion of several published fault injection and reliability experiments for ANNs [10]- [15] and SNNs [16], [20], [22], [23], [33], [35], [36] is that not all faults are equal. A large number of faults are either completely masked or they induce a negligible drop in the network classification accuracy.…”

“…This is because software and hardware implementations of neural networks closely match together [10]. Recent fault injection experiments in AI hardware accelerators [10]- [16] have shown that there is some inherent fault tolerance since many faults are benign, i.e., they are masked before their effect propagates to the output layer or they produce an output change that is tolerable. This is expected based on the analogy with biological neural networks which have remarkable fault tolerance capabilities.…”

Prediction of the Impact of Approximate Computing on Spiking Neural Networks via Interval Arithmetic

“…Furthermore, the proposed IA model can be generalized for other applications with complex computations, for example, NN designs, as long as we can model the network computation flow. For instance, our model can be generalized to help in accelerator design for convolutional SNNs, like those in previous works [18], [19].…”

Dependability of Alternative Computing Paradigms for Machine Learning: hype or hope?

“…All these problems can lead to performance degradation or operational failures, which in turn can have important consequences, especially for safety-critical systems [11]- [14]. It is thus crucial to determine the reliability of ML applications implemented leveraging on emerging computing paradigms, especially when they are deployed in safety-critical and mission-critical applications, such as robotics, aerospace, smart healthcare, and autonomous driving.…”