The Utility Problem in Case-Based Reasoning

摘要: Case-based reasoning systems may suffer from the utility problem, which occurs when knowledge learned in an attempt to improve a system's performance degrades instead. One of primary causes problem is slowdown conventional memories as number stored items increases. Unrestricted learning algorithms can swamp their memory system, causing system slow down more than average speedup provided by individual rules. Massive parallelism often offered solution this problem. However, most theoretical parallel models indicate that solutions fail scale up large sizes, and hardware implementations across wide class machines technologies back these predictions. Failing creation ideal concurrent-write random access machine, only lies coping strategies, such restricting extremely high or amount searched. provides excellent framework for implementation testing range methods policies with