Title: Datasize-Based Confidence Measure for a Learning Agent
Author: Wojciech Jamroga, Parlevink Group, University of Twente, Netherlands

In this paper a confidence measure is considered for an agent who tries to keep a probabilistic model of her environment of action. The measure is meant to capture only one factor of the agent's doubt -- namely, the issue whether the agent has been able to collect a sufficient number of observations. In this case stability of the agent's current knowledge may give some clue about the trust she can put in the model -- indeed, some researchers from the field of probability theory suggested that such confidence should be based on the variance of the model (over time).

In the paper two different measures are proposed, both based on aggregate variance of the estimator provided by the learning process. The way the measures work is investigated through some simple experiments with simulated software agents. It turns out that an agent can benefit from using such measures as means for 'self-reflection'. The simulations suggest that the agent's confidence should reflect the deviation of the her knowledge from the reality. They also show that it can be sometimes captured using very simple methods: a measure proposed by Wang is tested in this context, and it works seldom worse than the variance-based measures, although it seems completely ad hoc and not well suited for this particular setting of experiments at the first sight.

Keywords: multiagent systems, confidence measure, uncertainty, machine learning, user modeling.

Computational Intelligence Group @ Technical University of Clausthal Home
Human Media Interaction Group @ University of Twente
Computer Science Group @ University of Gdansk Last modified 2002-10-25