Wilka Carvalho

Wilka Carvalho is a Research Fellow at Harvard University's Kempner Institute for the Study of Natural and Artificial Intelligence, where she works closely with Sam Gershman. Her research focuses on developing deep reinforcement learning theories for human learning and generalization.

Dr. Carvalho earned her Ph.D. at the University of Michigan, where she studied deep reinforcement learning with Satinder Singh, Honglak Lee, and Richard Lewis, and was supported by the NSF GRFP and Rackham Merit Fellowship. During her PhD, she spent significant time at DeepMind working with Murray Shanahan, Daniel Zoran, and Danilo Rezende. She completed an M.S. in Computer Science at USC working with Yan Liu on machine learning for healthcare, and earned her B.S. in Physics at Stony Brook University working with Axel Drees on computational nuclear physics.

Her research interests center on how predictions about future rewards drive human thoughts and behaviors:

• Deep reinforcement learning
• Human learning and generalization
• Computational cognitive science
• Artificial intelligence
• Neuroscience applications

Dr. Carvalho's recent publications examine how people generalize to new tasks via counterfactual simulation, building agents that coordinate with humans without human data, and computational approaches to naturalistic experimental paradigms. Her work bridges theoretical machine learning with practical cognitive science applications.

She has received prestigious fellowships including:

• National Science Foundation Graduate Research Fellowship Program (NSF GRFP)
• Rackham Merit Fellowship

Dr. Carvalho is highly active in the academic community, regularly presenting her work at leading institutions including Harvard, MIT, Stanford, UC Berkeley, Columbia, NYU, UCLA, and Caltech. She maintains valuable academic resources for graduate students through her PhD Resources website and is open to collaboration opportunities with researchers across neuroscience and cognitive science disciplines.