Research Interests: Probabilistic Neurosymblic Approaches, Human-Robot Interaction, Statistical Relational Learning, Interpretable Learning from Demonstration

Research Description: My research interests lie in bridging neural and symbolic approaches for sustainable human-robot interactions. Neural approaches have shown their superiority in many perception tasks. However, they do not possess the ability to reason over the environment in abstract terms. This makes it hard to envision a safe human-robot co-living space: the agents cannot communicate with their human counterparts or be subject to the same regulations. It is, therefore, imperative to build systems that bridge these two models to enable efficient cooperation whilst still supporting the accurate and noise-robust learning of neural networks.