Organizers:
- Royal C Ihuaenyi, Northeastern University
- Juner Zhu, Northeastern University
- Emma Lejeune, Boston University
- Adrian Buganza Tepole, Purdue University
Description:
Advanced engineered systems, structures, and materials are getting smart, resilient, integrated – and overall, increasingly complex. Typical examples include electrochemical energy storage systems, biological materials, deployable structures, and soft robotics. In these examples, an important source of complexity is the coupling of multiple physical effects, such as mechanics, chemical reactions, mass and heat transfer, etc. In addition, many of these examples require accurate modeling and prediction at multiple spatial and temporal scales. A large number of variables and degrees of freedom make it a daunting task to characterize, design, and manipulate the systems, structures, and materials. In the past, physics-based or first-principle-based theories have achieved great successes but gradually suffered from the “curse of dimensionality.” Recently, many data-driven approaches, particularly machine learning, have shown prominent advantages in dealing with such high-dimensional problems. In this symposium, we welcome applications of different types of data-driven approaches to solve real-world problems in order to trigger valuable discussions on this promising tool. At the same time, the greater scientific community has recognized that data-driven approaches are usually agnostic and prone to unphysical failure. Therefore, we particularly encourage investigations on this issue and attempts to combine data-driven approaches with physics-based theories.