Spencer Yandrofski

Spencer  Yandrofski
MIT MechE | Undergraduate Research and Innovation Scholar
Advisor: George Barbastathis
Department: Mechanical Engineering
Areas of Research: Computational Design, Optics
Years: 2022-2023
Research Project Title:

Laser Aperture Optimization with Machine Learning

abstract:The goal of this project is to employ generative adversarial networks in order to arrive more efficiently at innovative designs for lens assemblies. The large number of parameters involved in optical lens systems create a very broad and complex design space, which makes this a useful application of deep learning approaches that can precisely identify patterns and relationships in the underlying response function much more easily than traditional design approaches. My role in this project will be to make progress from simple lens simulations to development of a generative model capable of creating complete and novel lens arrays that outperform existing designs. This model will be able to take advantage of computational models to explore traditionally unused complex lens shapes.

My interest in this project stems from my overall passion for machine learning and the unique solutions it offers to challenging engineering and design problems like this one. I am fascinated by the process behind efficient exploration of a design space, especially using novel computational techniques to achieve unique designs, and this project is a deep dive into complex techniques that are very promising in the field of deep generative design.