Gene Lewis

Autonomous vehicles, almost by definition, hold promise to replace the driver in an increasing number of scenarios. However, this dream brings with it an enormous set of complexities; fundamentally, the vehicle must be capable of replacing the cognitive capabilities of a human that are ancillary to the control task. These tasks, among others, include obstacle detection, semantic scene understanding, and behavior prediction. Recent advances in machine learning have dramatically improved performance on these tasks, but at the cost of interpretability and lack of predictability. Therefore, there is a fundamental tension in autonomous vehicle design, requiring a balance of AI that works well with AI that is predictable. Gene Lewis, graduate student in Computer Science, is primarily interested in approaches to resolving this tension, ranging from sensor fusion to pedestrian modeling to control for reducing perception uncertainty. Gene hopes that his work will help enable AI to enter industry at a more rapid pace, while securing consumer trust in trained data-driven systems.

Gene is currently pursuing an M.S. in Computer Science from Stanford University, and holds a B.S. in Computer Science from the same.

Student's Graduation Year

2019

Publications

Value Sensitive Design for Autonomous Vehicle Motion Planning