Subhayan De, PhD

• Ph.D. – Civil Engineering, University of Southern California, CA, 2018
• M.S. – Electrical Engineering, University of Southern California, CA, 2016
• M.Eng. – Structural Engineering, Indian Institute of Science, India, 2013
• B.Eng. – Civil Engineering, Jadavpur University, India, 2011

Dr. Subhayan De is an Assistant Professor in the Department of Mechanical Engineering and heads the Uncertainty Quantification, Learning, Inference, and Design (UQLID) lab at NAU. Prior to joining NAU, he was a postdoctoral research associate in Aerospace Engineering Sciences at the University of Colorado Boulder. The main goal of Dr. De’s research is to establish new probabilistic data-driven paradigms to efficiently develop and validate models using machine learning tools that can be used for the design of multi-scale multi-functional structural systems and materials.

The UQLID lab at NAU currently has three active research projects:

(1) Machine Learning for Uncertainty Quantification: Recently, machine learning (ML)-assisted models, such as neural networks, capable of describing some of the complex physical phenomena with good accuracy and reasonable computational cost are being increasingly used in engineering applications. For exercises involving many realizations of the engineering systems (e.g., uncertainty quantification, design under uncertainty), these ML-assisted models can be exploited to develop physics-based surrogate models that are easy to evaluate once trained but at the same time accurate. However, these networks require a large dataset to train. In this research project, efficient training of neural networks using smaller datasets for applications to engineering problems are explored.

(2) Topology Optimization under Uncertainty: In topology optimization (TO), a critical precursor step of additive manufacturing, we try to optimally distribute materials inside the structure to satisfy some performance criteria. However, in the presence of uncertainty in material properties, loading, and boundary conditions, achieving a meaningful optimized design is computationally burdensome owing to a large number of optimization variables in TO. In this research project, we are developing efficient algorithms tailored for TO applications to address the effects of uncertainty in the design process.

(3) Data-driven Probabilistic Model Validation: Based on the philosophy advanced by the famous statistician George P. Box: “Essentially, all models are wrong, but some are useful,” in this research project, an efficient hierarchical probabilistic model validation framework is being developed to address the effects of uncertainty in modeling physical systems utilizing measurement datasets from experiments.