Resilience Measure
Resilience measures quantify a system's ability to withstand, adapt to, and recover from disruptions. Current research focuses on developing robust and accurate methods for predicting and assessing resilience across diverse systems, employing machine learning models (including neural networks, graph neural networks, and clustering algorithms) and generative approaches to address data scarcity and model complex interactions. These advancements are crucial for improving the reliability and safety of critical infrastructure (e.g., power grids, supply chains), enhancing community preparedness for disasters, and building more resilient AI systems.
Papers
Thales: Formulating and Estimating Architectural Vulnerability Factors for DNN Accelerators
Abhishek Tyagi, Yiming Gan, Shaoshan Liu, Bo Yu, Paul Whatmough, Yuhao Zhu
Resilience Evaluation of Entropy Regularized Logistic Networks with Probabilistic Cost
Koshi Oishi, Yota Hashizume, Tomohiko Jimbo, Hirotaka Kaji, Kenji Kashima