Main Menu (Mobile)- Block
Main Menu - Block
Current Research:
To forage in an uncertain and changing world, animals should possess various generalizable strategies enabling them to behave effectively under different circumstances. In order to gain a holistic understanding of such strategies, one needs to probe behaviors by exposing animals in a wide range of tasks beyond simple classical conditioning. Even to date with an increasing interest in more naturalistic settings—due to lacking a principled way in designing such complex tasks—most experiments were largely exploratory with little expectations about whether the tasks are relevant to animals or whether an effective strategy can be evoked and identified. We thus aim to design highly discriminative tasks that are suitable in probing complex behaviors via an enumeration approach. Armed with enumeration, we discover tasks that are out of reach by conventional handcrafting approaches. Moving forwards, we plan to apply these tasks in an appropriate order to iteratively uncover the cognitive architectures across different species—from fruit flies to humans.
Biography
I am interested in applying theoretical and computational approaches from physics, computer science, and other quantitative disciplines to the fields of neuroscience. I seek for novel normative frameworks at the intersection of reinforcement learning, efficient coding theory and Bayesian inference that enables us to study and discover novel behavioral strategies. Long term, I aim to apply enumeration approach for developing theories to understand the origin of robustness and generalizability that are ubiquitous in biology systems.