{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,22]],"date-time":"2026-04-22T18:41:04Z","timestamp":1776883264158,"version":"3.51.2"},"publisher-location":"New York, NY, USA","reference-count":50,"publisher":"ACM","license":[{"start":{"date-parts":[[2023,11,11]],"date-time":"2023-11-11T00:00:00Z","timestamp":1699660800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/www.acm.org\/publications\/policies\/copyright_policy#Background"}],"content-domain":{"domain":["dl.acm.org"],"crossmark-restriction":true},"short-container-title":[],"published-print":{"date-parts":[[2023,11,12]]},"DOI":"10.1145\/3581784.3607076","type":"proceedings-article","created":{"date-parts":[[2023,10,30]],"date-time":"2023-10-30T20:34:48Z","timestamp":1698698088000},"page":"1-14","update-policy":"https:\/\/doi.org\/10.1145\/crossmark-policy","source":"Crossref","is-referenced-by-count":18,"title":["Prodigy: Towards Unsupervised Anomaly Detection in Production HPC Systems"],"prefix":"10.1145","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3627-7311","authenticated-orcid":false,"given":"Burak","family":"Aksar","sequence":"first","affiliation":[{"name":"Boston University, Boston, United States of America"},{"name":"Sandia National Laboratories, BOSTON, United States of America"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-1097-549X","authenticated-orcid":false,"given":"Efe","family":"Sencan","sequence":"additional","affiliation":[{"name":"Boston University, Boston, United States of America"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9282-8977","authenticated-orcid":false,"given":"Benjamin","family":"Schwaller","sequence":"additional","affiliation":[{"name":"Sandia National Laboratories, Livermore, United States of America"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0000-9651-0299","authenticated-orcid":false,"given":"Omar","family":"Aaziz","sequence":"additional","affiliation":[{"name":"Sandia National Laboratories, Albuquerque, United States of America"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0008-3950-1626","authenticated-orcid":false,"given":"Vitus J.","family":"Leung","sequence":"additional","affiliation":[{"name":"Sandia National Laboratories, Albuquerque, United States of America"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-8605-5795","authenticated-orcid":false,"given":"Jim","family":"Brandt","sequence":"additional","affiliation":[{"name":"Sandia National Laboratories, Livermore, United States of America"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1704-3838","authenticated-orcid":false,"given":"Brian","family":"Kulis","sequence":"additional","affiliation":[{"name":"Boston University, Boston, United States of America"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5038-2682","authenticated-orcid":false,"given":"Manuel","family":"Egele","sequence":"additional","affiliation":[{"name":"Boston University, Boston, United States of America"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6554-088X","authenticated-orcid":false,"given":"Ayse K.","family":"Coskun","sequence":"additional","affiliation":[{"name":"Boston University, Boston, United States of America"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"320","published-online":{"date-parts":[[2023,11,11]]},"reference":[{"key":"e_1_3_2_2_1_1","doi-asserted-by":"publisher","DOI":"10.1109\/SC.2014.18"},{"key":"e_1_3_2_2_2_1","doi-asserted-by":"publisher","DOI":"10.1109\/CLUSTER.2015.71"},{"key":"e_1_3_2_2_3_1","doi-asserted-by":"publisher","DOI":"10.2172\/1891960"},{"key":"e_1_3_2_2_4_1","volume-title":"ALBADross: Active Learning Based Anomaly Diagnosis for Production HPC Systems. In International Conference on Cluster Computing (CLUSTER). IEEE, 369--380","author":"Aksar Burak","year":"2022","unstructured":"Burak Aksar, Efe Sencan, Benjamin Schwaller, Omar Aaziz, Vitus J Leung, Jim Brandt, Brian Kulis, and Ayse K Coskun. 2022. ALBADross: Active Learning Based Anomaly Diagnosis for Production HPC Systems. In International Conference on Cluster Computing (CLUSTER). IEEE, 369--380."},{"key":"e_1_3_2_2_5_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-78713-4_11"},{"key":"e_1_3_2_2_6_1","doi-asserted-by":"publisher","DOI":"10.24963\/ijcai.2019\/932"},{"key":"e_1_3_2_2_7_1","doi-asserted-by":"publisher","DOI":"10.1109\/ICAPAI49758.2021.9462056"},{"key":"e_1_3_2_2_8_1","doi-asserted-by":"publisher","DOI":"10.1145\/3337821.3337907"},{"key":"e_1_3_2_2_9_1","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3403392"},{"key":"e_1_3_2_2_10_1","volume-title":"CoRR abs\/2003.05991","author":"Bank Dor","year":"2020","unstructured":"Dor Bank, Noam Koenigstein, and Raja Giryes. 2020. Autoencoders. CoRR abs\/2003.05991 (2020). arXiv:2003.05991"},{"key":"e_1_3_2_2_11_1","unstructured":"Elisabeth Baseman Sean Blanchard Nathan DeBardeleben Amanda Bonnie and Adam Morrow. 2016. Interpretable anomaly detection for monitoring of high performance computing systems. In Outlier Definition Detection and Description on Demand Workshop at SIGKDD. ACM 1--27."},{"key":"e_1_3_2_2_12_1","volume-title":"Paul Lin, William McDoniel, Sean Miller, Roger P. Pawlowski, Edward Geoffrey Phillips, and Nathan V. Roberts.","author":"Bettencourt Matthew Tyler","year":"2019","unstructured":"Matthew Tyler Bettencourt, Sidney Shields, Kristian Beckwith, Keith Cartwright, Eric C Cyr, Richard Michael Jack Kramer, Paul Lin, William McDoniel, Sean Miller, Roger P. Pawlowski, Edward Geoffrey Phillips, and Nathan V. Roberts. 2019. EMPIRE: Sandia's Next Generation Plasma Tool? Kokkosifying EMPIRE-Fluid. (4 2019)."},{"key":"e_1_3_2_2_13_1","doi-asserted-by":"publisher","DOI":"10.1145\/2503210.2503247"},{"key":"e_1_3_2_2_14_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.engappai.2019.07.008"},{"key":"e_1_3_2_2_15_1","doi-asserted-by":"publisher","DOI":"10.1109\/AICAS.2019.8771527"},{"key":"e_1_3_2_2_16_1","volume-title":"Kevin Pedretti, Narate Taerat, and Ron Velarde.","author":"Brandt James M","year":"2015","unstructured":"James M Brandt, David DeBonis, Ann C Gentile, Jim Lujan, Cindy Martin, David J Martinez, Stephen Lecler Olivier, Kevin Pedretti, Narate Taerat, and Ron Velarde. 2015. Enabling Advanced Operational Analysis Through Multi-subsystem Data Integration on Trinity. Technical Report. Sandia National Lab. (SNL-CA), Livermore, CA (United States); Sandia National."},{"key":"e_1_3_2_2_17_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2018.03.067"},{"key":"e_1_3_2_2_18_1","doi-asserted-by":"publisher","DOI":"10.1145\/3126908.3126926"},{"key":"e_1_3_2_2_19_1","doi-asserted-by":"publisher","DOI":"10.1109\/NOMS.2016.7502990"},{"key":"e_1_3_2_2_20_1","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v35i5.16523"},{"key":"e_1_3_2_2_21_1","volume-title":"International parallel and distributed processing symposium","author":"Dorier Matthieu","unstructured":"Matthieu Dorier, Gabriel Antoniu, Rob Ross, Dries Kimpe, and Shadi Ibrahim. 2014. CALCioM: Mitigating I\/O interference in HPC systems through cross-application coordination. In International parallel and distributed processing symposium. IEEE, 155--164."},{"key":"e_1_3_2_2_22_1","volume-title":"Explaining deep classification of time-series data with learned prototypes. 2429","author":"Gee Alan H","year":"2019","unstructured":"Alan H Gee, Diego Garcia-Olano, Joydeep Ghosh, and David Paydarfar. 2019. Explaining deep classification of time-series data with learned prototypes. 2429 (2019), 15."},{"key":"e_1_3_2_2_23_1","doi-asserted-by":"publisher","DOI":"10.2307\/2346830"},{"key":"e_1_3_2_2_24_1","doi-asserted-by":"publisher","DOI":"10.1214\/ss\/1177009869"},{"key":"e_1_3_2_2_25_1","doi-asserted-by":"publisher","DOI":"10.1145\/3437963.3441815"},{"key":"e_1_3_2_2_26_1","doi-asserted-by":"publisher","DOI":"10.1109\/CLUSTER.2017.23"},{"key":"e_1_3_2_2_28_1","unstructured":"Scott M Lundberg and Su-In Lee. 2017. A Unified Approach to Interpreting Model Predictions. In NeurIPS 30. 4765--4774."},{"key":"e_1_3_2_2_29_1","doi-asserted-by":"publisher","DOI":"10.1145\/3149412.3149421"},{"key":"e_1_3_2_2_30_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.parco.2004.04.001"},{"key":"e_1_3_2_2_31_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.future.2022.12.001"},{"key":"e_1_3_2_2_32_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.parco.2022.102950"},{"key":"e_1_3_2_2_33_1","doi-asserted-by":"publisher","DOI":"10.1109\/Cluster48925.2021.00086"},{"key":"e_1_3_2_2_34_1","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-59851-8_18"},{"key":"e_1_3_2_2_35_1","doi-asserted-by":"publisher","DOI":"10.1080\/14786440009463897"},{"key":"e_1_3_2_2_36_1","doi-asserted-by":"publisher","DOI":"10.1016\/0169-2070(89)90066-6"},{"key":"e_1_3_2_2_37_1","volume-title":"Elements of multivariate time series analysis","author":"Reinsel Gregory C","unstructured":"Gregory C Reinsel. 2003. Elements of multivariate time series analysis. Springer Science & Business Media."},{"key":"e_1_3_2_2_38_1","doi-asserted-by":"publisher","DOI":"10.1145\/2939672.2939778"},{"key":"e_1_3_2_2_39_1","doi-asserted-by":"publisher","DOI":"10.1109\/34.730550"},{"key":"e_1_3_2_2_40_1","volume-title":"Markus Hagenbuchner, and Gabriele Monfardini.","author":"Scarselli Franco","year":"2008","unstructured":"Franco Scarselli, Marco Gori, Ah Chung Tsoi, Markus Hagenbuchner, and Gabriele Monfardini. 2008. The graph neural network model. IEEE transactions on neural networks 20, 1 (2008), 61--80."},{"key":"e_1_3_2_2_41_1","volume-title":"Towards A Rigorous Evaluation Of XAI Methods On Time Series. In IEEE\/CVF International Conference on Computer Vision Workshop (ICCVW). 4197--4201","author":"Schlegel U.","unstructured":"U. Schlegel, H. Arnout, M. El-Assady, D. Oelke, and D. A. Keim. 2019. Towards A Rigorous Evaluation Of XAI Methods On Time Series. In IEEE\/CVF International Conference on Computer Vision Workshop (ICCVW). 4197--4201."},{"key":"e_1_3_2_2_42_1","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevE.55.5443"},{"key":"e_1_3_2_2_43_1","doi-asserted-by":"publisher","DOI":"10.1109\/CLUSTER49012.2020.00062"},{"key":"e_1_3_2_2_44_1","unstructured":"Burr Settles. 2009. Active learning literature survey. (2009)."},{"key":"e_1_3_2_2_45_1","doi-asserted-by":"publisher","DOI":"10.1109\/IISWC.2005.1526010"},{"key":"e_1_3_2_2_46_1","doi-asserted-by":"publisher","DOI":"10.1177\/1094342014522573"},{"key":"e_1_3_2_2_47_1","doi-asserted-by":"publisher","DOI":"10.1109\/ICIIC.2010.31"},{"key":"e_1_3_2_2_48_1","doi-asserted-by":"publisher","DOI":"10.14778\/3514061.3514067"},{"key":"e_1_3_2_2_49_1","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2018.2870403"},{"key":"e_1_3_2_2_50_1","volume-title":"A review of recurrent neural networks: LSTM cells and network architectures. Neural computation 31, 7","author":"Yu Yong","year":"2019","unstructured":"Yong Yu, Xiaosheng Si, Changhua Hu, and Jianxun Zhang. 2019. A review of recurrent neural networks: LSTM cells and network architectures. Neural computation 31, 7 (2019), 1235--1270."},{"key":"e_1_3_2_2_51_1","doi-asserted-by":"publisher","DOI":"10.1109\/CLUSTER49012.2020.00026"}],"event":{"name":"SC '23: International Conference for High Performance Computing, Networking, Storage and Analysis","location":"Denver CO USA","acronym":"SC '23","sponsor":["SIGHPC ACM Special Interest Group on High Performance Computing, Special Interest Group on High Performance Computing","IEEE CS"]},"container-title":["Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis"],"original-title":[],"link":[{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3581784.3607076","content-type":"unspecified","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/dl.acm.org\/doi\/pdf\/10.1145\/3581784.3607076","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,17]],"date-time":"2025-06-17T16:36:23Z","timestamp":1750178183000},"score":1,"resource":{"primary":{"URL":"https:\/\/dl.acm.org\/doi\/10.1145\/3581784.3607076"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,11,11]]},"references-count":50,"alternative-id":["10.1145\/3581784.3607076","10.1145\/3581784"],"URL":"https:\/\/doi.org\/10.1145\/3581784.3607076","relation":{},"subject":[],"published":{"date-parts":[[2023,11,11]]},"assertion":[{"value":"2023-11-11","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}]}}