活动主题：学术讲座From texts to reliable products: Assessing and improving reputational reliability based on natural language processing and machine learning
活动类型：学术交流
举办单位：智慧民航科创中心
活动时间：2024-12-27 14:00-16:00
活动地点：科研一号楼3004
面向群体：全院师生
主讲嘉宾：
曾志国，巴黎萨克雷大学教授
Professor Zhiguo ZENG received the Ph.D. degree in systems engineering from Beihang university in 2016. After receiving his PhD, he joined CentraleSupélec, Université Paris-Saclay, and became a full professor in 2023. His research focuses on the characterization and modeling of the failure/repair/maintenance behavior of components, complex systems and their reliability, maintainability, prognostics, safety, vulnerability and security. Dr. ZENG is an author/co-author of more than 150 papers in highly recognized international journals and conferences. He is recognized as Top Scholar by ScholarGPS based on the his strong publication records and impact of his research. His research has been funded by important government funding agencies like ANR and ERC, and also important industrial companies like EDF, SNCF, Orange and GE Healthcare. He is editorial board member of International Journal of Data Analysis Techniques and Strategies, and the leading guest editor of the special issue on “Dependent failure modeling” of the journal Applied Science. He is the co-head of the master program “Risk, Resilience and Engineering Management” in Universite Paris Saclay, and the engineering degree program “Operation Research and Risk Analytics”.
内容摘要：
In this talk, we briefly introduce some of our recent work on how to leverage text data to extract reliability-related information and improve product reliability. Our first work concerns using customer review data for reputational reliability assessment. First, a dataset for failure detection from customer review data was created, along with a human benchmark to account for natural language ambiguity. Then, both traditional natural language processing techniques and more contemporary methods involving BERT variations were benchmarked on this task and the best developed model came to be a fine-tuned transformer, which reaches 88% balanced accuracy for failure detection, against an estimated human performance of 91%. Finally, the extracted failure was used to generate lifetime data to assess their reliability and mean-time-to-failure.
联系人：智慧民航科创中心，李颖异
