Lyle R
- Research Program Mentor
PhD candidate at Massachusetts Institute of Technology (MIT)
Expertise
Machine Learning, Computer Science, Design, Mechanics, Dynamics, Robotics
Bio
I am a current graduate student at the MIT Design Computation and Digital Engineering Lab. My research revolves around harnessing advancements in machine learning to enable AI-driven design. Through my research, I ultimately hope to empower design engineers around the world with AI-assisted tools to enable them to take on design challenges that would otherwise be insurmountable. Currently, I am developing Hierarchical cGANs with Performance-Aware Feedback as a means to address the generative design synthesis problem. During my undergraduate studies at the University of Illinois at Urbana-Champaign, I double majored in mechanical engineering and electrical engineering and minored in computer science. My curriculum and research interests were primarily focused around robotics, but I undertook research projects ranging from designing a Cation Intercalation Desalination cell to optimizing atomic deposition processes to developing a robotic hovercraft system for coordinated multimodal drone operations. During high school my favorite pastimes were math competitions (AMC/AIME/USAMO) and robotics (FIRST).Project ideas
Modeling Complex Phenomena using Machine Learning
In this project you will model a complex real-world phenomenon with trained regressive models. First, you will select a domain of interest and generate a dataset either through simulation, curating public information, or physical experimentation. A few domain ideas: • Predicting likes on a social media post • Estimating aerodynamic drag on an object • Estimating the rate of a chemical reaction • Predicting stock prices (Hard) • Predicting weather (Hard) The possibilities are endless, so pick something that excites you, but remember that you need a good way to generate your dataset! After generating a dataset, you will explore the data using unsupervised learning methods, then contrast performance for various regression models such as decision trees, neural networks and ensemble methods. Finally, you will study the behavior of the highest performing algorithms using tractability analyses, drawing connections to the real-world phenomena that may be driving these observations.
Machine/Product Design
In this project, you will design and model a product, machine or mechanism. Emphasis will be placed on proper design technique and rigorous mechanical and/or electrical design principles. Depending on the nature of the project, many particular considerations will be relevant including: • Design for Manufacturability/Assemblability • Electronics/printed circuit board design • Power Transmission • Material selection and processing • Structural Analysis • Fluid dynamics • Sensing and Logic • Aesthetics and ease of use Prototyping, design iteration, and actual fabrication may be possible, depending on available resources. This is intended to be an avenue to pursue a creative concept that you hope to realize. I will leverage my expertise from countless design cycles across numerous engineering disciplines including the aforementioned to help you turn your concept into reality. Pick something that you are passionate about and let’s make it happen!