Probability Distributions.- Linear Models for Regression.- Linear Models for Classification.- Neural Networks.- Kernel Methods.- Sparse Kernel Machines.- Graphical Models.- Mixture Models and EM.- Approximate Inference.- Sampling Methods.- Continuous Latent Variables.- Sequential Data.- Combining Models.

Pattern Recognition and Machine Learning

Monte Carlo tree search (MCTS) is a recently proposed search method that combines the precision of tree search with the generality of random sampling. It has received considerable interest due to its spectacular success in the difficult problem of computer Go, but has also proved beneficial in a range of other domains. This paper is a survey of the literature to date, intended to provide a snapshot of the state of the art after the first five years of MCTS research. We outline the core algorithm's derivation, impart some structure on the many variations and enhancements that have been proposed, and summarize the results from the key game and nongame domains to which MCTS methods have been applied. A number of open research questions indicate that the field is ripe for future work.

/pdf/a-survey-of-monte-carlo-tree-search-methods-2mj9olkdni.pdf

A Survey of Monte Carlo Tree Search Methods

https://iridia.ulb.ac.be/IridiaTrSeries/link/IridiaTr2011-004.pdf

The irace package: Iterated racing for automatic algorithm configuration

Statistics: Methods and Applications.

This paper gives an overview of the open source project Potassco, the Potsdam Answer Set Solving Collection, bundling tools for Answer Set Programming developed at the University of Potsdam.

Potassco: The Potsdam Answer Set Solving Collection

We propose a portfolio-based solving approach to Answer Set Programming (ASP). Our approach is homogeneous in considering several configurations of the ASP solver clasp. The selection among the configurations is realized via Support Vector Regression. The resulting portfolio-based solver claspfolio regularly outperforms clasp's default configuration as well as manual tuning.

/pdf/a-portfolio-solver-for-answer-set-programming-preliminary-4k3jjt78ob.pdf

A portfolio solver for answer set programming: preliminary report

We present the General Game Playing system Centurio. Centurio is a Java-based player featuring different strategies based on Monte Carlo Tree Search extended by techniques borrowed from Upper Confidence bounds applied to Trees as well as Answer Set Programming (for single-player games). Centurio’s Monte Carlo Tree Search is accomplished in a massively parallel way by means of multi-threading as well as cluster-computing. Another major feature of Centurio is its compilation of game descriptions, states, and state manipulations into Java, yielding an edge over existing Prolog-based approaches. Centurio is open source software freely available via the web.

/pdf/centurio-a-general-game-player-parallel-java-and-asp-based-s7iw0mv484.pdf

Centurio, a General Game Player: Parallel, Java- and ASP-based

We investigate the composition of representative benchmark sets for evaluating and improving the performance of robust Boolean constraint solvers in the context of satisfiability testing and answer set programming. Starting from an analysis of current practice, we isolate a set of desiderata for guiding the development of a parametrized benchmark selection algorithm. Our algorithm samples a benchmark set from a larger base set or distribution comprising a large variety of instances. This is done fully automatically, in a way that carefully calibrates instance hardness and instance similarity. We demonstrate the usefulness of this approach by means of empirical results showing that optimizing solvers on the benchmark sets produced by our method leads to better configurations than obtained based on the much larger, original sets.

/pdf/robust-benchmark-set-selection-for-boolean-constraint-55cd62a0h5.pdf

Robust Benchmark Set Selection for Boolean Constraint Solvers

Automated configuration procedures play an increasingly prominent role in realising the performance potential inherent in highly parametric solvers for a wide range of computationally challenging problems. However, these configuration procedures have difficulties when dealing with inhomogenous instance sets, where the relative difficulty of problem instances varies between configurations of the given parametric algorithm. In the literature, instance set homogeneity has been assessed using a qualitative, visual criterion based on heat maps. Here, we introduce two quantitative measures of homogeneity and empirically demonstrate these to be consistent with the earlier qualitative criterion. We also show that according to our measures, homogeneity increases when partitioning instance sets by means of clustering based on observed runtimes, and that the performance of a prominent automatic algorithm configurator increases on the resulting, more homogenous subsets.

/pdf/quantifying-homogeneity-of-instance-sets-for-algorithm-1721nyrzl2.pdf

Marius Schneider

Papers

Potassco: The Potsdam Answer Set Solving Collection

A portfolio solver for answer set programming: preliminary report

Centurio, a General Game Player: Parallel, Java- and ASP-based

Robust Benchmark Set Selection for Boolean Constraint Solvers

Quantifying homogeneity of instance sets for algorithm configuration