Active Learning on Non-Stationary Functions - Robotics Institute Carnegie Mellon University

Active Learning on Non-Stationary Functions

Conference Paper, Proceedings of AAAI '95 Fall Symposium on Active Learning, November, 1995

Abstract

We consider the problem of active learning on a function that varies with time. A common approach is to employ a learning algorithm that reduces the contribution of data points as they become older. With this method an active learner will decide to resample regions of the space where all the data has become fairly old. We propose an alternative that calls for the learner to retain all data. As the function varies in time, it may stray away from a certain mapping, but may also return to it at some point in the future. Our algorithm reviews its historical data to see if the current mapping is similar to one held previously and re-uses the old data points when this occurs. Our results show that this can provide better approximation of time-varying functions and can free an active learner from having to resample regions of the space where historical data has become relevant again.

BibTeX

@conference{Schneider-1995-16254,
author = {Jeff Schneider},
title = {Active Learning on Non-Stationary Functions},
booktitle = {Proceedings of AAAI '95 Fall Symposium on Active Learning},
year = {1995},
month = {November},
}