Data driven techniques have raised continuous interest in a variety of industries, including manufacturing, transportation, education and energy etc. Despite tremendous progress, the currenting data driven approaches are fully data driven, not capable of coping with mass of monitoring and detection data, and with little domain knowledge in processing, diagnosis and prediction. However, there are often fewer samples, due to robust manufacturing technologies and continuous data monitoring early warning applied in safety-critical complex equipment and assets. As a sequence, current data driven approaches often have poor generalization performance and given untrustworthy diagnostic decisions for operators. Therefore, it is considerably urgent to investigate and explore innovative paradigms for big data processing, synergize data driven techniques with domain-specific knowledge and capture the dynamic characteristic and failure propagation with such as partial differential equations, causal graphs. These paradigms have the potential to substantially enhance the flexibility and trustworthiness of the data driven approaches, and further strengthen their both safety and reliability.
Chair: Jian Wang, Southwest Jiaotong University, China (firstname.lastname@example.org)
Co-Chair: Lei Kou, Shandong Academy of Sciences, China (email@example.com)
Co-Chair: Junhe Wan, Shandong Academy of Sciences, China (firstname.lastname@example.org)
● Submission Method:
● Topics of interest include (but are not limited to:
Data driven methods;
Big data processing;
Data-driven prognostics and health management;
Trustworthy decision-making support;
Data driven technologies for system safety, reliability, and security;
Data augmentation learning;
Few shot learning;
Graph augmentation learning;
Internet of Things.