Transductive Confidence Machine for Active Learning

Shen Shyang Ho, Harry Wechsler

Research output: Contribution to conferencePaperpeer-review

20 Scopus citations

Abstract

This paper describes a novel active learning strategy using universal p-value measures of confidence based on algorithmic randomness, and transductive inference. The early stopping criteria for active learning is based on the bias-variance trade-off for classification. This corresponds to that learning instance when the boundary bias becomes positive, and requires one to switch from active to random selection of learning examples. The sign for the boundary bias and the increase in the classification error are two manifestations of the same phenomena, i.e., over-training. The experimental results presented show the feasibility and usefulness of our novel approach using a non-separable two-class classification problem. Our hybrid learning strategy achieves competitive performance against standard nearest neighbor methods using much fewer training examples.

Original languageEnglish (US)
Pages1435-1440
Number of pages6
StatePublished - Sep 24 2003
Externally publishedYes
EventInternational Joint Conference on Neural Networks 2003 - Portland, OR, United States
Duration: Jul 20 2003Jul 24 2003

Other

OtherInternational Joint Conference on Neural Networks 2003
Country/TerritoryUnited States
CityPortland, OR
Period7/20/037/24/03

All Science Journal Classification (ASJC) codes

  • Software

Fingerprint

Dive into the research topics of 'Transductive Confidence Machine for Active Learning'. Together they form a unique fingerprint.

Cite this