Stochastic search algorithms are black-box optimizer of an objective function. They have recently gained a lot of attention in operations research, machine learning and policy search of robot motor skills due to their ease of use and their generality. However, when the task or objective function slightly changes, many stochastic search algorithms require complete re-learning in order to adapt thesolution to the new objective function or the new context. As such, we consider the contextual stochastic search paradigm. Here, we want to find good parameter vectors for multiple related tasks, where each task is described by a continuous context vector. Hence, the objective function might change slightly for each parameter vector evaluation. In this paper, we investigate a contextual stochastic search algorithm known as Contextual Relative Entropy Policy Search (CREPS), an information-theoretic algorithm that can learn from multiple tasks simultaneously. We show the application of CREPS for simulated robotic tasks.

Extensive-form games with imperfect recall are an important model of dynamic games where the players forget previously known information. Often, imperfect recall games are the result of an abstraction algorithm that simplifies a large game with perfect recall. Unfortunately, solving an imperfect recall game has fundamental problems since a Nash equilibrium does not have to exist. Alternatively, we can seek maxmin strategies that guarantee an expected outcome. The only existing algorithm computing maxmin strategies in imperfect recall games, however, requires approximating a bilinear program that is proportional to the size of the game and thus has a limited scalability. We propose a novel algorithm for computing maxmin strategies that combines this approximate algorithm with an incremental strategy-generation technique designed previously for extensive-form games with perfect recall. Experimental evaluation shows that the novel algorithm builds only a fraction of the game tree and improves the scalability by several orders of magnitude. Finally, we demonstrate that our algorithm can solve an abstracted variant of a large game faster compared to the algorithms operating on the unabstracted perfect-recall variant.

Kidney exchanges are organized markets where patients swap willing but incompatible donors. In the last decade, kidney exchanges grew from small and regional to large and national---and soon, international.  This growth results in more lives saved, but exacerbates the empirical hardness of the NP-complete problem of optimally matching patients to donors.  State-of-the-art matching engines use integer programming techniques to clear fielded kidney exchanges, but these methods must be tailored to specific models and objective functions, and may fail to scale to larger exchanges. In this paper, we observe that if the kidney exchange compatibility graph can be encoded by a constant number of patient and donor attributes, the clearing problem is solvable in polynomial time. We give necessary and sufficient conditions for losslessly shrinking the representation of an arbitrary compatibility graph. Then, using real compatibility graphs from the UNOS US-wide kidney exchange, we show how many attributes are needed to encode real graphs. The experiments show that, indeed, small numbers of attributes suffice.

If you're looking to play around with robotics, Lego's Mindstorms EV3 is a great way to get started. So is the ultra-versatile Raspberry Pi. Combining the two to create a Rubik's Cube-solving robot? That sounds like a good time to us! The Lego bricks take care of the physical moves required to solve the puzzle.

A typical viral marketing model identifies influential users in a social network to maximize a single product adoption assuming unlimited user attention, campaign budgets, and time. In reality, multiple products need campaigns, users have limited attention, convincing users incurs costs, and advertisers have limited budgets and expect the adoptions to be maximized soon. Facing these user, monetary, and timing constraints, we formulate the problem as a submodular maximization task in a continuous-time diffusion model under the intersection of a matroid and multiple knapsack constraints. We propose a randomized algorithm estimating the user influence in a network ($|\mathcal{V}|$ nodes, $|\mathcal{E}|$ edges) to an accuracy of $\epsilon$ with $n=\mathcal{O}(1/\epsilon^2)$ randomizations and $\tilde{\mathcal{O}}(n|\mathcal{E}|+n|\mathcal{V}|)$ computations. By exploiting the influence estimation algorithm as a subroutine, we develop an adaptive threshold greedy algorithm achieving an approximation factor $k_a/(2+2 k)$ of the optimal when $k_a$ out of the $k$ knapsack constraints are active. Extensive experiments on networks of millions of nodes demonstrate that the proposed algorithms achieve the state-of-the-art in terms of effectiveness and scalability.

Finding the right classifier to use for your data can be hard. Once you have chosen a classifier, tuning all of the parameters to get the best results is tedious and time consuming. Even after all of your hard work, you may have chosen the wrong classifier to begin with. Hyperopt-sklearn provides a solution to this problem. Any search algorithm available in hyperopt can be used to drive the estimator.

This paper presents a particular model of heuristic search as a path-finding problem in a directed graph. A class of graph-searching procedures is described which uses a heuristic function to guide search. Heuristic functions are estimates of the number of edges that remain to be traversed in reaching a goal node. A number of theoretical results for this model, and the intuition for these results, are presented. They relate the efficiency of search to the accuracy of the heuristic function.

Abstract: Real-time situated agents, including characters in real-time computer games, often do not know the terrain in advance but automatically observe it within a certain range around them. They have to interleave planning with movement to make planning tractable when moving autonomously to user-specified coordinates. Planning faces real-time requirements since it is important that the agents be responsive to the commands of the users and move smoothly. In this paper, we compare two fast search methods for this task that speed up planning in different ways, namely real-time heuristic search (LRTA*) and incremental heuristic search (D* Lite), resulting in the first comparison of real-time and incremental heuristic search in the literature. We characterize when to choose which search method, depending on the kind of terrain and the planning objective.

Board Games Research Group: develops high performance search algorithms and game playing programs such as Fuego, the first Go program to beat a top human player in 9x9 Go. Games Research Group: engages in the design, analysis, and implementation of artificial intelligence technology that is suitable for use in high-performance game-playing programs. Intelligent Reasoning Critiquing and Learning (IRCL) Group: conducts Artificial Intelligence research on real-time heuristic search, interactive story-telling and cognitive modeling. Our recent applications have been with video games. We have on-going collaborations with the Department of Psychology, UBC Okanagan, Reykjavik University and Disney Research.

In the popular imagination, chess isn't like a spelling bee or Trivial Pursuit, a competition to see who can hold the most facts in memory and consult them quickly. In chess, as in the arts and sciences, there is plenty of room for beauty, subtlety, and deep originality. Chess requires brilliant thinking, supposedly the one feat that would be–forever–beyond the reach of any computer. But for a decade, human beings have had to live with the fact that one of our species' most celebrated intellectual summits–the title of world chess champion–has to be shared with a machine, Deep Blue, which beat Garry Kasparov in a highly publicized match in 1997. What lessons could be gleaned from this shocking upset?