“…Current state-of-the-art functional methods for detecting faults that may affect AI-powered devices are often based on Software Test Libraries (STLs), a set of assembly programs that can excite and detect faults in the underlying computing device by applying carefully crafted test patterns. However, since STLs require a significant manual effort to develop and may affect the performance of the AI task [22], alternative solutions have been proposed that leverage directly the computations of AI models, such as CNNs. In particular, they involve constructing suitable input data that, when fed to the CNN, can excite faults simply by flowing through the layers of the network and propagate them up to the output layer.…”
This study presents a comparative examination of state-of-the-art resiliency approaches of Convolutional, Spiking, and Photonic neural networks (CNNs, SNNs, PNNs), their fault and error models, and the main fault tolerance techniques.
“…Current state-of-the-art functional methods for detecting faults that may affect AI-powered devices are often based on Software Test Libraries (STLs), a set of assembly programs that can excite and detect faults in the underlying computing device by applying carefully crafted test patterns. However, since STLs require a significant manual effort to develop and may affect the performance of the AI task [22], alternative solutions have been proposed that leverage directly the computations of AI models, such as CNNs. In particular, they involve constructing suitable input data that, when fed to the CNN, can excite faults simply by flowing through the layers of the network and propagate them up to the output layer.…”
This study presents a comparative examination of state-of-the-art resiliency approaches of Convolutional, Spiking, and Photonic neural networks (CNNs, SNNs, PNNs), their fault and error models, and the main fault tolerance techniques.
“…Algorithmic-based error detection and correction methods using checksum arithmetic are discussed in [13], [26]- [29]. On-line test methods are proposed in [30] based on Software Test Libraries (STL), in [31] based on a simplified metric of dynamic power consumption, and in [32] based on encrypting weights in the memory with an encryption algorithm that spreads single bitflips extending them to multiple bit-flips and checking if the padding bytes used for the encryption to work properly are correctly decrypted.…”
Section: Prior Art On Testing Ai Hardware Acceleratorsmentioning
We address the problem of testing Artificial Intelligence (AI) hardware accelerators implementing Spiking Neural Networks (SNNs). We define a metric to quickly rank available samples for training and testing based on their fault detection capability. The metric measures the inter-class spike count difference of a sample for the fault-free design. In particular, each sample is assigned a score equal to the spike count difference between the first two top classes. The hypothesis is that samples with small scores achieve high fault coverage because they are prone to misclassification, i.e., a small perturbation in the network due to a fault will result in these samples being misclassififed with high probability. We show that the proposed metric correlates with the per-sample fault coverage and that retaining a set of high-ranked samples in the order of ten achieves near perfect fault coverage for critical faults that affect the SNN accuracy. The proposed test generation approach is demonstrated on two SNNs modelled in Python and on actual neuromorphic hardware. We discuss fault modeling and perform an analysis to reduce the fault space so as to speed up test generation time.
“…3) Software-based: In [157], on-line test strategies based on Software Test Libraries (STL) are proposed for embedded systems running ANN applications. STL is composed of selftest routines that are executed during boot-time or run-time.…”
Section: On-line Test 1) Atpg and Functional Testingmentioning
confidence: 99%
“…These solutions bring in new challenges for testing. For instance, in-memory solutions may require understanding and creation Physical-aware DFT [136] DFT overhead Function-aware DFT [134], [135], [145] Complete DFT solutions for large heterogeneous systems Functional test generation [127], [146]- [156] Memory-hungry designs On-line test [157]- [165] New market demands (i.e., 0 DPPM for automotive, on-line test)…”
In recent years, there has been an expedited trend in embracing bold and radical innovation of computer architectures, aiming at the continuation of computing performance improvement despite the slowed-down physical device scaling. One new frontier in this field focuses on Artificial Intelligence (AI) hardware. While functionality of AI hardware still remains the main focus, testability and dependability of these new architectures need to be addressed before the mainstream adoption. This survey paper covers the state-of-the-art in research and development of dependability and testability solutions for AI hardware including digital or analog implementations of Artificial Neural Networks (ANNs) and Spiking Neural Networks (SNNs), used in accelerators and neuromorphic designs. Trends, challenges and perspectives are also discussed in this paper.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.