Searching with a Corrupted Heuristic

Abstract

Memory-based heuristics are a popular and effective class of admissible heuristic functions. However, corruptions to memory they use may cause these heuristics to become inadmissible. Corruption can be caused by the physical environment due to radiation and network errors, or it can be introduced voluntarily in order to decrease energy consumption. We introduce memory error correction schemes that do not require additional memory and exploit knowledge about the behavior of consistent heuristics. This is in contrast with error correcting code approaches which can limit the amount of corruption but at the cost of additional energy and memory consumption. Search algorithms using our methods are guaranteed to find a solution if one exists and its suboptimality is bounded. Moreover, our methods are resilient to any number of memory errors that may occur. An experimental evaluation is also provided to demonstrate the applicability of our approach.

Optimal Search with Neural Networks: Challenges and Approaches

“…For instance, Shah et al (2020) describes a method to learn a heuristic with at most error, but in empirical discussions it is conceded that this is only with high probability -under the assumption that the heuristic function can be precisely learned. High probability is not good enough for A*, although it could work for other algorithms (Ernandes and Gori 2004;Stern, Felner, and Holte 2011;Lelis et al 2016). Similarly, other recent work (Yonetani et al 2021) proposed that a neural architecture could learn to do suboptimal pathfinding 'better' than A* and Weighted A*.…”