This paper describes an unsupervised machine-learning method for detecting single-event transient (SET) radiation effects in microelectronics and classifying areas within circuits sensitive to radiation effects. A neural network creates latent spaces for classifying transient phenomena and minimizes loss through the mean squared error function. Differences in the loss within the latent space allow the categorizing of major groups of unique SET responses. A data set including 100,000 transient waveforms obtained through carrier generation based on sub-bandgap two-photon absorption (TPA) laser-induced single event effects is used to demonstrate the data clustering technique. After initial model training, the testing data set is placed in the latent space and used to train a K-Means model, which allows for automatic clustering of related waveforms. The clustering technique is validated by illustrating the data clusters on a two-dimensional spatial mapping of sensitive regions obtained from the TPA experiments. The method can quickly identify radiation-induced anomalies, providing near real-time assessment of measured data. It provides a computationally and temporally cheap way to analyze and group SET anomalies without examining the entire data set or using a supervised technique requiring labeled training data. The model can ingest any new incoming data and predict the disruption’s originating location.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Classification of Microelectronics Radiation Effects Using Unsupervised Machine Learning


    Beteiligte:
    Peyton, Trevor (Autor:in) / Carpenter, Jake (Autor:in) / Reising, Donald (Autor:in) / Loveless, Daniel (Autor:in)


    Erscheinungsdatum :

    02.03.2024


    Format / Umfang :

    17507853 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Radiation effects on microelectronics in space

    Srour, J.R. / McGarrity, J.M. | Tema Archiv | 1988



    Engine Usage Analysis Using Unsupervised Machine Learning

    Cerqueira, S. / Morel, H. | British Library Conference Proceedings | 2023


    Microelectronics

    C. Husson | NTIS | 1964


    A roadmap for NASA's radiation effects research in emerging microelectronics and photonics

    LaBel, K.A. / Barnes, C.E. / Marshall, P.W. et al. | IEEE | 2000