Blossom Metevier

I am a graduate student in the Autonomous Learning Lab at the University of Massachusetts, Amherst, where I am co-advised by Phil Thomas and Lee Spector. Before attending UMass, I was a member of the Force Projection Sector at APL. I completed my undergraduate degree in Computer Science from the University of Maryland, Baltimore County.

My research area is in reinforcement learning and genetic programming. In particular, I am interested in the development of self-organizing agents whose resulting structure allows for robust performance in complex environments. I believe that some important problems in reinforcement learning, including task generalization and coping with high-dimensional state and/or action space, can be alleviated using evolutionary methods. I am also interested in nonconvex optimization and blackbox optimization algorithms. Currently, I am working on applications of fairness in the contextual bandit setting with high probability guarantees.

Apart from the grad student grind, I enjoy running and reading. I love Hemmingway's theory of omission and Ursula Guin's alternative worlds, though in general I tend towards fantasy. I am also a fan of the DC universe, specifically the Teen Titans, and I follow a number of Japanese comics. I workout often and like spending time outdoors when it is warm. Many activities are appealing to me, including running, badminton, lifting, and kayaking.

Contact information: bmetevier [at] cs [dot] umass [dot] edu