Improvement of Variational Autoencoder Based Test Escape Detection through Image Transformation

• Item type: SIG Technical Reports(1)
• 公開日: 2019-11-06
• タイトル: Improvement of Variational Autoencoder Based Test Escape Detection through Image Transformation
• タイトル / 言語: en
• タイトル: Improvement of Variational Autoencoder Based Test Escape Detection through Image Transformation
• 言語: eng
• 資源タイプ / 資源タイプ識別子: http://purl.org/coar/resource_type/c_18gh
• 資源タイプ: technical report
• 著者所属: Graduate School of Science and Technology
• 著者所属: Graduate School of Science and Technology
• 著者所属: Renesas Electronics Corporation
• 著者所属: Graduate School of Science and Technology
• 著者所属(英): en
• 著者所属(英): Graduate School of Science and Technology
• 著者所属(英): en
• 著者所属(英): Graduate School of Science and Technology
• 著者所属(英): en
• 著者所属(英): Renesas Electronics Corporation
• 著者所属(英): en
• 著者所属(英): Graduate School of Science and Technology
• 著者名: Remain, Chicoix; Michlhiro, Shintani; Kouichi, Kumaki; Michiko, Inoue
• 著者名(英): Remain, Chicoix; Michlhiro, Shintani; Kouichi, Kumaki; Michiko, Inoue
• 論文抄録 / 内容記述タイプ: Other
• 論文抄録 / 内容記述: In testing of large scale integration (LSI) circuit, test escape detection using machine-learning algorithms has been attracted attention. More particular, variational autoencoder (VAE) can detect faults that could not be detected by conventional test methods by learning features of fault-free LSIs as random variables. Meanwhile, computer vision is one of the most successful fields in machine learning applications. Inspired by the fact, in this paper, the LSI test data is converted into images, and they are used to improve the detection performance of the outlier detection using a VAE. Through experiments, the proposed approach achieves approximately 2x higher outlier detectability than conventional work.
• 論文抄録(英) / 内容記述タイプ: Other
• 論文抄録(英) / 内容記述: In testing of large scale integration (LSI) circuit, test escape detection using machine-learning algorithms has been attracted attention. More particular, variational autoencoder (VAE) can detect faults that could not be detected by conventional test methods by learning features of fault-free LSIs as random variables. Meanwhile, computer vision is one of the most successful fields in machine learning applications. Inspired by the fact, in this paper, the LSI test data is converted into images, and they are used to improve the detection performance of the outlier detection using a VAE. Through experiments, the proposed approach achieves approximately 2x higher outlier detectability than conventional work.
• 書誌レコードID / 収録物識別子タイプ: NCID
• 書誌レコードID / 収録物識別子: AA11451459
• 書誌情報: 研究報告システムとLSIの設計技術（SLDM）; 巻 2019-SLDM-189, 号 3, p. 1-6, 発行日 2019-11-06
• ISSN / 収録物識別子タイプ: ISSN
• ISSN / 収録物識別子: 2188-8639
• Notice: SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc.
• 出版者 / 言語: ja
• 出版者: 情報処理学会