Enhancing AI System Resiliency: Formulation and Guarantee for LSTM Resilience Based on Control Theory
This provides a theoretical foundation for rigorous quality assurance in safety-critical AI applications, though it appears incremental in applying control theory to LSTMs.
This paper tackles the problem of guaranteeing LSTM network resilience in control systems by introducing 'recovery time' as a new metric and deriving a data-independent upper bound on it using refined δISS theory, with experimental validation on simple models demonstrating effectiveness.
This paper proposes a novel theoretical framework for guaranteeing and evaluating the resilience of long short-term memory (LSTM) networks in control systems. We introduce "recovery time" as a new metric of resilience in order to quantify the time required for an LSTM to return to its normal state after anomalous inputs. By mathematically refining incremental input-to-state stability ($δ$ISS) theory for LSTM, we derive a practical data-independent upper bound on recovery time. This upper bound gives us resilience-aware training. Experimental validation on simple models demonstrates the effectiveness of our resilience estimation and control methods, enhancing a foundation for rigorous quality assurance in safety-critical AI applications.