Machine common sense remains a broad, potentially unbounded problem in AI. Our focus is to move toward AI systems that can develop common-sense reasoning similar to humans to detect anomalies. In particular, we study the problem of detecting the violation of expectations when object appearance or motion dynamics change from...
This dissertation delves into understanding, characterizing, and addressing dataset shift in deep learning, a pervasive issue for deployed machine learning systems. Integral aspects of the problem are examined: We start with the use of counterfactual explanations in order to characterize the behavior of deep reinforcement learning agents in visual input...
In a scenario where multiple Unmanned Aerial Vehicles (UAVs) run by different operators are conducting their independent missions in an urban environment, it is important that the limited airspace is shared in a fair manner. Clarity and uniformity in communicating mission specifications to a central controller is also crucial. This...