Overview

I aim to make smarter robots that can safely operate with and around humans.

Towards this end, my research focuses on developing explainable machine learning models that allow robots to operate with full transparency. Using principled approaches for learning models that balance explainability and accuracy, these robots are able to explain how and why they perform actions.

Explanations are important for two reasons:

1. We can understand whether behavior is based on sound reasoning, which allows us to decide if the robot can be trusted to act safely and correctly.
2. We can perform targeted interventions when this is not the case, allowing us to correct behavior in situ.

My work has enabled robots to hug, collaboratively manipulate objects, drive alongside humans, and more. Examples of my work are shown in the video below.