Enhancing AI System Resiliency: Formulation and Guarantee for LSTM Resilience Based on Control Theory

This research proposes methods for formulating and guaranteeing the resilience of long short-term memory (LSTM) networks, which can serve as a key technology in AI system quality assurance. We introduce a novel methodology applying incremental input-to-state stability ($\delta$ISS) to mathematically define and evaluate the resilience of LSTM against input perturbations. Key achievements include the development of a data-independent evaluation method and the demonstration of resilience control through adjustments to training parameters. This research presents concrete solutions to AI quality assurance from a control theory perspective, which can advance AI applications in control systems.

Spike Accumulation Forwarding for Effective Training of Spiking Neural Networks

In this article, we propose a new paradigm for training spiking neural networks (SNNs), spike accumulation forwarding (SAF). It is known that SNNs are energy-efficient but difficult to train. Consequently, many researchers have proposed various methods to solve this problem, among which online training through time (OTTT) is a method that allows inferring at each time step while suppressing the memory cost. However, to compute efficiently on GPUs, OTTT requires operations with spike trains and weighted summation of spike trains during forwarding. In addition, OTTT has shown a relationship with the Spike Representation, an alternative training method, though theoretical agreement with Spike Representation has yet to be proven. Our proposed method can solve these problems; namely, SAF can halve the number of operations during the forward process, and it can be theoretically proven that SAF is consistent with the Spike Representation and OTTT, respectively. Furthermore, we confirmed the above contents through experiments and showed that it is possible to reduce memory and training time while maintaining accuracy.