I am entering my third year as a PhD candidate at the University of Michigan in the AI lab. I am co-advised by Satinder Singh and Edmund Durfee.

I am interested in sequential decision making problems in artificial intelligence in which the agent must actively acquire additional information in order to better accomplish the goals of its designer. The agent must select which information to acquire while simultaneously behaving well according to its designer's standards, which means that if the information comes at a cost, the agent must determine whether it is worth obtaining at all.

In semi-autonomous agents, a human operator is available in some form to help the agent act on his/her behalf. Here, the human operator is the provider of information, and pestering the human is costly. Previously I studied how the agent can utilize costly help from the operator, for several different types of help the human can provide, when the agent's knowledge of its environment is represented as a Bayesian Markov Decision Processes.