Biblio

In recent years, Deep Neural Networks (DNNs) have made inroads into a number of applications involving pattern recognition - from facial recognition to self-driving cars. Some of these applications, such as self-driving cars, have real-time requirements, where specialized DNN hardware accelerators help meet those requirements. Since DNN execution time is dominated by convolution, Multiply-and-Accumulate (MAC) units are at the heart of these accelerators. As hardware accelerators push the performance limits with strict power constraints, reliability is often compromised. In particular, power-constrained DNN accelerators are more vulnerable to transient and intermittent hardware faults due to particle hits, manufacturing variations, and fluctuations in power supply voltage and temperature. Methods such as hardware replication have been used to deal with these reliability problems in the past. Unfortunately, the duplication approach is untenable in a power constrained environment. This paper introduces a low-cost error-resiliency scheme that targets MAC units employed in conventional DNN accelerators. We evaluate the reliability improvements from the proposed architecture using a set of 6 CNNs over varying bit error rates (BER) and demonstrate that our proposed solution can achieve more than 99% of fault coverage with a 5-bits arithmetic code, complying with the ASIL-D level of ISO26262 standards with a negligible area and power overhead. Additionally, we evaluate the proposed detection mechanism coupled with a word masking correction scheme, demonstrating no loss of accuracy up to a BER of 10-2.

Authorship attribution, an NLP problem where anonymous text is matched to its author, has important, cross-disciplinary applications, particularly those concerning cyber-defense. Our research examines the degree of sensitivity that attention-based models have to adversarial perturbations. We ask, what is the minimal amount of change necessary to maximally confuse a transformer model? In our investigation we examine a balanced subset of emails from the Enron email dataset, calculating the performance of our model before and after email signatures have been perturbed. Results show that the model's performance changed significantly in the absence of a signature, indicating the importance of email signatures in email authorship detection. Furthermore, we show that these models rely on signatures for shorter emails much more than for longer emails. We also indicate that additional research is necessary to investigate stylometric features and adversarial training to further improve classification model robustness.