We present PROCAB, an efficient method for Probabilistically Reasoning from Observed Context-Aware Behavior. It models the context-dependent utilities and underlying reasons that people take different actions. The model generalizes to unseen situations and scales to incorporate rich contextual information. We train our model using the route preferences of 25 taxi drivers demonstrated in over 100,000 miles of collected data, and demonstrate the performance of our model by inferring: (1) decision at next intersection, (2) route to known destination, and (3) destination given partially traveled route.

Modeling the purposeful behavior of imperfect agents from a small number of observations is a challenging task. When restricted to the single-agent decision-theoretic setting, inverse optimal control techniques assume that observed behavior is an approximately optimal solution to an unknown decision problem. These techniques learn a utility function that explains the example behavior and can then be used to accurately predict or imitate future behavior in similar observed or unobserved situations. In this work, we consider similar tasks in competitive and cooperative multi-agent domains. Here, unlike single-agent settings, a player cannot myopically maximize its reward — it must speculate on how the other agents may act to influence the game’s outcome. Employing the game-theoretic notion of regret and the principle of maximum entropy, we introduce a technique for predicting and generalizing behavior, as well as recovering a reward function in these domains. 1.

Community-authored content, such as location specific reviews, offers a wealth of information about virtually every imaginable location today. In this work, we process Yelp‘s community-authored reviews to identify a set of potential activities that are supported by the location reviewed. Using 14 test locations we show that the majority of the 40 most common results per location (determined by verb-noun pair frequency) are actual activities supported by their respective locations, achieving a mean precision of up to 79.3%. Although the number of reviews authored for a location has a strong influence on precision, we are able to achieve a precision up to 29.5 % when processing only the first 50 reviews, increasing to 45.7 % and 57.3 % for the first 100 and 200 reviews, respectively. In addition, we present two context-aware services that leverage location-based activity information on a city scale that is accessible through a Web service we developed supporting multiple cities in North America. Author Keywords activity, community-authored content, location, reviews