Rare class discovery based on active learning

Jingrui He, Jaime Carbonell

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

In machine learning, the new-class discovery problem remains an open challenge, especially for emergent rare classes. However, the challenge is of crucial importance for applications such as detecting new financial fraud patterns, new viral mutations and new network malware, most of which 'hide' among vast volumes of normal data and observations. This paper focuses on a new approach, based on local-topology density estimation, applicable to discovering examples of the rare classes rapidly, despite non-separability with the majority class(es). The new method, called ALICE, and its variant MALICE, are shown effective both theoretically and empirically in outperforming other methods in the literature, both on challenging synthetic data and on real data sets.

Original languageEnglish (US)
Title of host publication10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008
StatePublished - 2008
Externally publishedYes
Event10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008 - Fort Lauderdale, FL, United States
Duration: Jan 2 2008Jan 4 2008

Other

Other10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008
CountryUnited States
CityFort Lauderdale, FL
Period1/2/081/4/08

Fingerprint

Active Learning
Learning systems
Topology
Malware
Density Estimation
Synthetic Data
Machine Learning
Mutation
Class
Problem-Based Learning

ASJC Scopus subject areas

  • Artificial Intelligence
  • Applied Mathematics

Cite this

He, J., & Carbonell, J. (2008). Rare class discovery based on active learning. In 10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008

Rare class discovery based on active learning. / He, Jingrui; Carbonell, Jaime.

10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008. 2008.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

He, J & Carbonell, J 2008, Rare class discovery based on active learning. in 10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008. 10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008, Fort Lauderdale, FL, United States, 1/2/08.
He J, Carbonell J. Rare class discovery based on active learning. In 10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008. 2008
He, Jingrui ; Carbonell, Jaime. / Rare class discovery based on active learning. 10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008. 2008.
@inproceedings{506ddf7e69b64b3f98ae5e8c0a2bfa89,
title = "Rare class discovery based on active learning",
abstract = "In machine learning, the new-class discovery problem remains an open challenge, especially for emergent rare classes. However, the challenge is of crucial importance for applications such as detecting new financial fraud patterns, new viral mutations and new network malware, most of which 'hide' among vast volumes of normal data and observations. This paper focuses on a new approach, based on local-topology density estimation, applicable to discovering examples of the rare classes rapidly, despite non-separability with the majority class(es). The new method, called ALICE, and its variant MALICE, are shown effective both theoretically and empirically in outperforming other methods in the literature, both on challenging synthetic data and on real data sets.",
author = "Jingrui He and Jaime Carbonell",
year = "2008",
language = "English (US)",
booktitle = "10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008",

}

TY - GEN

T1 - Rare class discovery based on active learning

AU - He, Jingrui

AU - Carbonell, Jaime

PY - 2008

Y1 - 2008

N2 - In machine learning, the new-class discovery problem remains an open challenge, especially for emergent rare classes. However, the challenge is of crucial importance for applications such as detecting new financial fraud patterns, new viral mutations and new network malware, most of which 'hide' among vast volumes of normal data and observations. This paper focuses on a new approach, based on local-topology density estimation, applicable to discovering examples of the rare classes rapidly, despite non-separability with the majority class(es). The new method, called ALICE, and its variant MALICE, are shown effective both theoretically and empirically in outperforming other methods in the literature, both on challenging synthetic data and on real data sets.

AB - In machine learning, the new-class discovery problem remains an open challenge, especially for emergent rare classes. However, the challenge is of crucial importance for applications such as detecting new financial fraud patterns, new viral mutations and new network malware, most of which 'hide' among vast volumes of normal data and observations. This paper focuses on a new approach, based on local-topology density estimation, applicable to discovering examples of the rare classes rapidly, despite non-separability with the majority class(es). The new method, called ALICE, and its variant MALICE, are shown effective both theoretically and empirically in outperforming other methods in the literature, both on challenging synthetic data and on real data sets.

UR - http://www.scopus.com/inward/record.url?scp=84864590022&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84864590022&partnerID=8YFLogxK

M3 - Conference contribution

BT - 10th International Symposium on Artificial Intelligence and Mathematics, ISAIM 2008

ER -