CASUS Institute Seminar, Prof. Dr. Sinnu Susan Thomas, Digital University Kerala (Indian Institute of Information Technology and Management-Kerala)
With the rise of artificial intelligence, the industry has more than ever been searching for new ways to further improve the design of new products. Harvesting of the data has never been cheaper and the computing powers are always increasing and in a similar context, we focus on the design of automatic MacPherson suspension architecture for the automotive sector. Engineering design is traditionally performed by hand: an expert makes design proposals based on past experience, and these proposals are then tested for compliance with certain target specifications. Testing for compliance is performed first by computer simulation using what is called a discipline model. Such a model can be implemented by a finite element analysis, multibody systems approach, etc. Designs passing this simulation are then considered for physical prototyping. The overall process may take months, and is a significant cost in practice.
In this talk, I would talk about a Gaussian Processes particularly Bayesian Optimization system that partially automates this process by directly optimizing compliance with the target specification with respect to the design parameters. I would discuss different convergence criteria to attain an optimal solution and then would talk about the dataset shift learning paradigm for the biased data generated using the discipline models.
This talk would also cover the applications of Bayesian Optimization for controller synthesis and stellar parameters estimation.