Machine learning in predicting and explaining failure using class-imbalanced fab data

H. Belyavin; B. Lerner

Machine learning in predicting and explaining failure using class-imbalanced fab data

H. Belyavin, B. Lerner

Department of Industrial Engineering and Management

Ben-Gurion University of the Negev

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

Defective lots due to an ion-implantation fab tool producing flash memories can be detected only six weeks after implantation, leading to material waste and financial loss. Machine learning (ML) is used to predict and explain tool failure. For training the ML model, a small and heavily imbalanced database is used with only 8% records of defective lots. It undermines knowledge representation and accurate prediction of defective lots. A methodology is presented that includes correlation analysis, discretization, down sampling, and model optimization using a performance measure, which is imbalance-insensitive. The methodology improves the accuracy of the ML models in classifying defective lots. The models also identify variables related to ion suppression and extraction voltages and currents and thereby may explain the failure.

Original language	American English
Title of host publication	21st International Conference on Production Research
Subtitle of host publication	Innovation in Product and Production, ICPR 2011 - Conference Proceedings
Editors	Tobias Krause, Dieter Spath, Rolf Ilg
ISBN (Electronic)	9783839602935
State	Published - 1 Jan 2011
Event	21st International Conference on Production Research: Innovation in Product and Production, ICPR 2011 - Stuttgart, Germany Duration: 31 Jul 2011 → 4 Aug 2011

Conference

Conference	21st International Conference on Production Research: Innovation in Product and Production, ICPR 2011
Country/Territory	Germany
City	Stuttgart
Period	31/07/11 → 4/08/11

Keywords

Bayesian networks
Failure prediction
Machine learning
Neural networks
Semiconductor manufacturing

All Science Journal Classification (ASJC) codes

Control and Systems Engineering
Computer Science Applications
Industrial and Manufacturing Engineering

Cite this

@inproceedings{42b924b7a0fe4342b22126911881b69f,

title = "Machine learning in predicting and explaining failure using class-imbalanced fab data",

abstract = "Defective lots due to an ion-implantation fab tool producing flash memories can be detected only six weeks after implantation, leading to material waste and financial loss. Machine learning (ML) is used to predict and explain tool failure. For training the ML model, a small and heavily imbalanced database is used with only 8% records of defective lots. It undermines knowledge representation and accurate prediction of defective lots. A methodology is presented that includes correlation analysis, discretization, down sampling, and model optimization using a performance measure, which is imbalance-insensitive. The methodology improves the accuracy of the ML models in classifying defective lots. The models also identify variables related to ion suppression and extraction voltages and currents and thereby may explain the failure.",

keywords = "Bayesian networks, Failure prediction, Machine learning, Neural networks, Semiconductor manufacturing",

author = "H. Belyavin and B. Lerner",

year = "2011",

month = jan,

day = "1",

language = "American English",

editor = "Tobias Krause and Dieter Spath and Rolf Ilg",

booktitle = "21st International Conference on Production Research",

note = "21st International Conference on Production Research: Innovation in Product and Production, ICPR 2011 ; Conference date: 31-07-2011 Through 04-08-2011",

}

Belyavin, H & Lerner, B 2011, Machine learning in predicting and explaining failure using class-imbalanced fab data. in T Krause, D Spath & R Ilg (eds), 21st International Conference on Production Research: Innovation in Product and Production, ICPR 2011 - Conference Proceedings. 21st International Conference on Production Research: Innovation in Product and Production, ICPR 2011, Stuttgart, Germany, 31/07/11.

Machine learning in predicting and explaining failure using class-imbalanced fab data. / Belyavin, H.; Lerner, B.
21st International Conference on Production Research: Innovation in Product and Production, ICPR 2011 - Conference Proceedings. ed. / Tobias Krause; Dieter Spath; Rolf Ilg. 2011.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Machine learning in predicting and explaining failure using class-imbalanced fab data

AU - Belyavin, H.

AU - Lerner, B.

PY - 2011/1/1

Y1 - 2011/1/1

N2 - Defective lots due to an ion-implantation fab tool producing flash memories can be detected only six weeks after implantation, leading to material waste and financial loss. Machine learning (ML) is used to predict and explain tool failure. For training the ML model, a small and heavily imbalanced database is used with only 8% records of defective lots. It undermines knowledge representation and accurate prediction of defective lots. A methodology is presented that includes correlation analysis, discretization, down sampling, and model optimization using a performance measure, which is imbalance-insensitive. The methodology improves the accuracy of the ML models in classifying defective lots. The models also identify variables related to ion suppression and extraction voltages and currents and thereby may explain the failure.

AB - Defective lots due to an ion-implantation fab tool producing flash memories can be detected only six weeks after implantation, leading to material waste and financial loss. Machine learning (ML) is used to predict and explain tool failure. For training the ML model, a small and heavily imbalanced database is used with only 8% records of defective lots. It undermines knowledge representation and accurate prediction of defective lots. A methodology is presented that includes correlation analysis, discretization, down sampling, and model optimization using a performance measure, which is imbalance-insensitive. The methodology improves the accuracy of the ML models in classifying defective lots. The models also identify variables related to ion suppression and extraction voltages and currents and thereby may explain the failure.

KW - Bayesian networks

KW - Failure prediction

KW - Machine learning

KW - Neural networks

KW - Semiconductor manufacturing

UR - http://www.scopus.com/inward/record.url?scp=84923478696&partnerID=8YFLogxK

M3 - Conference contribution

BT - 21st International Conference on Production Research

A2 - Krause, Tobias

A2 - Spath, Dieter

A2 - Ilg, Rolf

T2 - 21st International Conference on Production Research: Innovation in Product and Production, ICPR 2011

Y2 - 31 July 2011 through 4 August 2011

ER -

Machine learning in predicting and explaining failure using class-imbalanced fab data

Abstract

Conference

Keywords

All Science Journal Classification (ASJC) codes

Other files and links

Fingerprint

Cite this