The Utility Problem in Case-Based Reasoning

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 performance instead. One of the primary causes of the utility problem is the slowdown of conventional memories as the number of stored items increases. Unrestricted learning algorithms can swamp their memory system, causing the system to slow down more than the average speedup provided by individual learned rules.

Massive parallelism is often offered as a solution to this problem. However, most theoretical parallel models indicate that parallel solutions to the utility problem fail to scale up to large problem sizes, and hardware implementations across a wide class of machines and technologies back up these predictions.

Failing the creation of an ideal concurrent-write parallel random access machine, the only solution to the utility problem lies in a number of coping strategies, such as restricting learning to extremely high utility items or restricting the amount of memory searched. Case-based reasoning provides an excellent framework for the implementation and testing of a wide range of methods and policies for coping with the utility problem.

Read the paper:

The Utility Problem in Case-Based Reasoning

by Anthony Francis, Ashwin Ram

AAAI-93 Workshop on Case-Based Reasoning, Washington, DC, July 1993
www.cc.gatech.edu/faculty/ashwin/papers/er-93-08.pdf
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: