With smart-phones becoming increasingly commonplace, there has been a subsequent surge in applications that continuously track the location of users. However, serious privacy concerns arise as people start to widely adopt these applications. Users will need to maintain policies to determine under which circumstances to share their location. Specifying these policies however, is a cumbersome task, suggesting that machine learning might be helpful. In this paper, we present a user-controllable method for learning location sharing policies. We use a classifier based on multivariate Gaussian mixtures that is suitably modified so as to restrict the evolution of the underlying policy to favor incremental and therefore human-understandable changes as new data arrives. We evaluate the model on real location-sharing policies collected from a live location-sharing social network, and we show that our method can learn policies in a user-controllable setting that are just as accurate as policies that do not evolve incrementally. Additionally, we highlight the strength of the generative modeling approach we take, by showing how our model easily extends to the semi-supervised setting.

We consider the problem of routing electric vehicles (EV) in the most energy-efficient way within a road network taking into account both their limited energy supply as well as their ability to recuperate energy. Employing a classical result by Johnson and an observation about Dijkstra under non-constant edge costs we obtain O(n log n +m) query time after a O(nm) preprocessing phase for any road network graph whose edge costs represent energy consumption or recuperation.If the energy recuperation is height induced in a very natural way,the preprocessing phase can even be omitted. We then adapt a technique for speeding-up (unconstrained) shortest path queries to our scenario to achieve a speed-up of another factor of around 20. Our results drastically improve upon the recent results in (Artmeier et al. 2010) and allow for route planning of EVs in an instant even on large networks.

Recently, there has been a significant interest in studying security games to provide tools for addressing resource allocation problems in security applications. Patrolling security games (PSGs) constitute a special class of security games wherein the resources are mobile. One of the most relevant open problems in security games is the design of scalable algorithms to tackle realistic scenarios. While the literature mainly focuses on heuristics and decomposition techniques (e.g., double oracle), in this paper we provide, to the best of our knowledge, the first study on the use of abstractions in security games (specifically for PSGs) to design scalable algorithms. We define some classes of abstractions and we provide parametric algorithms to automatically generate abstractions. We show that abstractions allow one to relax the constraint of patrolling strategies' Markovianity (customary in PSGs) and to solve large game instances. We additionally pose the problem to search for the optimal abstraction and we develop an anytime algorithm to find it.

We present a novel approach for identifying exact and approximate behavioral equivalence between models of agents. This is significant because both decision making and game play in multiagent settings must contend with behavioral models of other agents in order to predict their actions. One approach that reduces the complexity of the model space is to group models that are behaviorally equivalent. Identifying equivalence between models requires solving them and comparing entire policy trees. Because the trees grow exponentially with the horizon, our approach is to focus on partial policy trees for comparison and determining the distance between updated beliefs at the leaves of the trees. We propose a principled way to determine how much of the policy trees to consider, which trades off solution quality for efficiency. We investigate this approach in the context of the interactive dynamic influence diagram and evaluate its performance.

Dual-decomposition (DD) methods are quickly becoming important tools for estimating the minimum energy state of a graphical model. DD methods decompose a complex model into a collection of simpler subproblems that can be solved exactly (such as trees), that in combination provide upper and lower bounds on the exact solution. Subproblem choice can play a major role: larger subproblems tend to improve the bound more per iteration, while smaller subproblems enable highly parallel solvers and can benefit from re-using past solutions when there are few changes between iterations. We propose an algorithm that can balance many of these aspects to speed up convergence. Our method uses a cluster tree data structure that has been proposed for adaptive exact inference tasks, and we apply it in this paper to dual-decomposition approximate inference. This approach allows us to process large subproblems to improve the bounds at each iteration, while allowing a high degree of parallelizability and taking advantage of subproblems with sparse updates. For both synthetic inputs and a real-world stereo matching problem, we demonstrate that our algorithm is able to achieve significant improvement in convergence time.

When to make a decision is a key question in decision making problems characterized by uncertainty. In this paper we deal with decision making in environments where the information arrives dynamically. We address the tradeoff between waiting and stopping strategies. On the one hand, waiting to obtain more information reduces the uncertainty, but it comes with a cost. On the other hand, stopping and making a decision based on an expected utility, decreases the cost of waiting, but the decision is made based on uncertain information. In this paper, we prove that computing the optimal time to make a decision that guarantees the optimal utility is NP-hard. We propose a pessimistic approximation that guarantees an optimal decision when the recommendation is to wait. We empirically evaluate our algorithm and show that the quality of the decision is near-optimal and much faster than the optimal algorithm.

Recent domain-determinization techniques have been very successful in many probabilistic planning problems. We claim that traditional heuristic MDP algorithms have been unsuccessful due mostly to the lack of efficient heuristics in structured domains. Previous attempts like mGPT used classical planning heuristics to an all-outcome determinization of MDPs without discount factor; yet, discounted optimization is required to solve problems with potential dead-ends. We propose a general extension of classical planning heuristics to goal-oriented discounted MDPs, in order to overcome this flaw. We apply our theoretical analysis to the well-known classical planning heuristics Hmax and Hadd, and prove that the extended Hmax is admissible. We plugged our extended heuristics to popular graph-based (Improved-LAO*, LRTDP, LDFS) and ADD-based (sLAO*, sRTDP) MDP algorithms: experimental evaluations highlight competitive results compared with the winners of previous competitions (FF-Replan, FPG, RFF), and show that our discounted heuristics solve more problems than non-discounted ones, with better criteria values. As for classical planning, the extended Hadd outperforms the extended Hmax on most problems.

Previous research in Artificial Intelligence has identified the possibility of simplifying planning problems via the identification and exploitation of symmetries. We advance the state of the art in algorithms that exploit symmetry in planning problems by generalizing previous approaches, and applying symmetry reductions to state-based planners. We suggest several algorithms for symmetry exploitation in state-based search, but also provide a comprehensive view through which additional algorithms can be developed and fine-tuned. We evaluate our approach to symmetry exploitation on instances from previous planning competitions, and demonstrate that our algorithms significantly improve the solution time of instances with symmetries.

Symbolic search with BDDs has shown remarkable performance for cost-optimal deterministic planning by exploiting a succinct representation and exploration of state sets. In this paper we enhance BDD-based planning by applying a combination of domain-independent search techniques: the optimization of the variable ordering in the BDD by approximating the linear arrangement problem, pattern selection for improved construction of search heuristics in form of symbolic partial pattern databases, and a decision procedure for the amount of bidirection in the symbolic search process.

Let h be a heuristic function selected for expansions when GBFS with the FF heuristic that estimates the distance to a goal from a node n. GBFS (Hoffmann and Nebel 2001) solves a planning problem. The selects the best node n with the smallest h(n) in the open list horizontal axis indicates each expansion of the best node that maintains nodes that have been generated but have not n in the open list and the vertical axis represents n's corresponding been expanded yet. It then expands n to generate n's successors, heuristic value for that expansion. Circles, the and saves these successors in the open list, unless triangle, and diamond represent expanding nodes that are they have been previously added to the open list.