Group-Exclusive Feature Group Lasso and Applications to Automatic Sensor Selection for Virtual Metrology in Semiconductor Manufacturing

Jeongsub Choi, Youngdoo Son, Jihoon Kang

Research output: Contribution to journalArticlepeer-review

Abstract

Group lasso is a regularization widely used for feature group selection with sparsity at a group level in machine learning. Training a model with the group lasso regularization, however, leads to the selection of all the groups together that are closely related to each other although their features are useful to predict a target. In this study, we propose a new regularization, group-exclusive group lasso, for automatic exclusive feature group selection. The proposed regularization aims to enforce exclusive sparsity at an inter-group level, discouraging the coincident selection of the feature groups that are group-level correlated and share predictive powers toward the targets. The proposed method aims at higher group sparsity for selecting salient feature groups only, and is applied to neural networks. We evaluate the proposed regularization in neural networks on synthetic datasets and a real-life case for virtual metrology with automatic sensor selection in semiconductor manufacturing.

Original languageEnglish
Pages (from-to)505-517
Number of pages13
JournalIEEE Transactions on Semiconductor Manufacturing
Volume37
Issue number4
DOIs
StatePublished - 2024

Keywords

  • Group exclusivity
  • group sparsity
  • regularization
  • sensor selection
  • virtual metrology

Fingerprint

Dive into the research topics of 'Group-Exclusive Feature Group Lasso and Applications to Automatic Sensor Selection for Virtual Metrology in Semiconductor Manufacturing'. Together they form a unique fingerprint.

Cite this