Data-Driven Sparse Modeling and Intelligent Soft Sensing for Complex Industrial Processes

Chair:	Co-Chair:
Yanjun Liu  Jiangnan University, China	Siyu Liu Zhejiang Normal University, China

Key Words: Sparse Optimization, Data-Driven Identification, Soft Sensing, State Estimation, Process Monitoring

Information: The rapid advancement of industrial intelligence has generated massive datasets, yet extracting meaningful low-dimensional features from high-dimensional, noisy environments remains a significant challenge. This session focuses on the intersection of sparse learning, system identification, and soft sensor technology. We explore how sparsity-promoting techniques (such as $L_1$-norm regularization) enhance the interpretability and robustness of system models. By leveraging sparse learning, researchers can identify the core structures of complex dynamical systems and develop efficient soft sensors for real-time estimation of hard-to-measure variables. The session discusses innovations in sparse optimization, nonlinear identification, and intelligent sensing deployment in volatile industrial settings. We invite contributions that bridge the gap between theoretical modeling and practical challenges—such as variable selection and multi-rate fusion—to improve monitoring reliability across sectors like chemical processing, energy management, and advanced manufacturing.

Topics of interest include but are not limited to: 

• Identification of Nonlinear Dynamical Systems by Sparse learning
• Robust Soft Sensing in the Presence of Missing Data and Outliers
• Industrial Applications of Sparse Learning in Energy, Chemical, and Manufacturing Systems
• Soft Sensor-Based State Estimation and Feedback in Predictive Control
• Robust MPC Design under Sparse Uncertainty and Variable Selection

Submission Deadline: June 30, 2026 (the second round)