- J. Miekisz, M. Matuszak, J. Poleszczuk,
Stochastic Stability in Three-Player Games with Time Delays, Dynamic Games and Applications, 1-10, DOI 10.1007/s13235-014-0115-1, May 2014.
- P. Kaminska, M. Matuszak,
Image Segmentation by Locally Specified Multi-coloured Polygonal Markov Fields, Proceedings of the 2nd International Conference on Intelligent Systems and Image Processing, 2014.
- M. Nowicki, M. Matuszak, A.B. Kwiatkowska, M.M. Syslo, P. Bala,
Teaching secondary school students programming using distance learning. A case study,WCCE 2013, Vol. 2, pp. 246-254, 2013. Abstract.Regional Computer Science Course is a novel approach to increasing programming skills among secondary school students in Poland. It is fully based on distance learning and individual work of students after regular school hours. In this project the achievements of students were monitored on-line through weekly programming contests. The standardized tests have been developed to monitor students' skills. In result large number of students increased their programming skills. In this paper we present idea of the project as well as its advantages and pitfalls.
- K. Choromanski, M. Matuszak, J. Miekisz,
Scale-Free Graph with Preferential Attachment and Evolving Internal Vertex Structure,Journal of Statistical Physics, Volume 151, Issue 6, pp. 1175-1183, June 2013. Abstract.We extend the classical Barabási-Albert preferential attachment procedure to graphs with internal vertex structure given by weights of vertices. In our model, weight dynamics depends on the current vertex degree distribution and the preferential attachment procedure takes into account both weights and degrees of vertices. We prove that such a coupled dynamics leads to scale-free graphs with exponents depending on parameters of the weight dynamics.
- M. Matuszak, J. Miekisz,
Stochastic Techniques in Influence Diagrams for Learning Bayesian Network Structure,ICANN 2012, Part I, LNCS 7552, pp. 33-40, 2012. Abstract.The problem of learning Bayesian network structure is well known to be NP–hard. It is therefore very important to develop efficient approximation techniques. We introduce an algorithm that within the framework of influence diagrams translates the structure learning problem into the strategy optimisation problem, for which we apply the Chen’s self–annealing stochastic optimisation algorithm. The effectiveness of our method has been tested on computer–generated examples.
- M. Matuszak, J. Miekisz, T. Schreiber,
Solving Ramified Optimal Transport Problem in the Bayesian Influence Diagram Framework,ICAISC 2012, Part II, LNCS 7268, pp. 582-590, 2012. Abstract.The goal of ramified optimal transport is to find an optimal transport path between two given probability measures. One measure can be identified with a source while the other one with a target. The problem is well known to be NP-hard. Within the framework of Bayesian networks we develop an algorithm for solving a ramified optimal transport problem based on the decision strategy optimisation technique that utilises self-annealing ideas of Chen-style stochastic optimisation. Resulting transport paths are represented in the form of tree-shaped structures. The effectiveness of the algorithm has been tested on computer-generated examples.
- M. Matuszak, T. Schreiber,
Locally specified polygonal Markov fields for image segmentation,Mathematical Methods for Signal and Image Analysis and Representation, Series: Computational Imaging and Vision, Vol. 41, Florack, L.; Duits, R.; Jongbloed, G.; Lieshout, M.-C. van; Davies, L. (Eds.), ISBN 978-1-4471-2352-1, 2012 Abstract.We introduce a class of polygonal Markov fields driven by local activity functions. Whereas the local rather than global nature of the field specification ensures substantial additional flexibility for statistical applications in comparison to classical polygonal fields, we show that a number of simulation algorithms and graphical constructions, as developed in our previous joint work with M.N.M. van Lieshout and R. Kluszczynski, carry over to this more general framework. Moreover, we provide explicit formulae for the partition function of the model, which directly implies the availability of closed form expressions for the corresponding likelihood functions. Within the framework of this theory we develop an image segmentation algorithm based on Markovian optimisation dynamics combining the simulated annealing ideas with those of Chen-style stochastic optimisation, in which successive segmentation updates are carried out simultaneously with adaptive optimisation of the local activity functions.
- M. Matuszak, J. Miekisz, T. Schreiber,
Smooth Conditional Transition Paths in Dynamical Gaussian Networks,KI 2011: Advances in Artificial Intelligence, LNAI 7006, pp. 204-215, 2011. Abstract.We propose an algorithm for determining optimal transition paths between given configurations of systems consisting ofmany objects. It is based on the Principle of Least Action and variational equations for Freidlin–Wentzell action functionals in Gaussian networks set-up.We use our method to construct a system controlling motion and redeployment between unit’s formations. Another application of the algorithm allows a realistic transformation between two sequences of character animations in a virtual environment. The efficiency of the algorithm has been evaluated in a simple sandbox environment implemented with the use of the NVIDIA CUDA technology.
- M. Matuszak, T. Schreiber,
GPU Accelerated Smooth Formation Redeployment in Multiagent Environment,MASYW 2010, Institute of Computer Science, Polish Academy of Sciences, pp. 92-100, Warsaw 2011. Abstract.The problem of determining optimal formation reorganization in a given time plays important role in many fields from aeronautics like satellites or missiles reconfigurations to entertainment such as controlling agents in computer games. We introduce an algorithm for discovering optimal transition paths between given configurations. Crucial role in presented method plays efficient computation of the matrix exponential. We recalled one of the best algorithms in solving that task - the scaling and squaringmethod combinedwith Pade approximants. To test the efficiency of the algorithm we have implemented a simple sandbox environment with use of the NVIDIA CUDA technology.
- M. Matuszak, T. Schreiber,
A new stochastic algorithm for strategy optimisation in Bayesian influence diagrams,ICAISC 2010, Part II, LNAI 6114, pp. 574-581, 2010. Abstract.The problem of solving general Bayesian influence diagrams is well known to be NP-complete, whence looking for efficient approximate stochastic techniques yielding suboptimal solutions in reasonable time is well justified. The purpose of this paper is to propose a new stochastic algorithm for strategy optimisation in Bayesian influence diagrams. The underlying idea is an extension of that presented in  by Chen who developed a self-annealing algorithm for optimal tour generation in traveling salesman problems (TSP). Our algorithm generates optimal decision strategies by iterative self-annealing reinforced search procedure, gradually acquiring new information while driven by information already acquired. The effectiveness of our method has been tested on computer-generated examples.
- J. Matulewski, M. Pakulski, D. Borycki, B. Bialy, P. Peplowski, M. Matuszak, D. Szlag, D. Urbanski,
Visual C++. Gotowe rozwiazania dla programistow Windows (eng. Visual C++ - Practical Solutions for Windows Developers),ISBN: 978-83-246-1928-3, 2010 Abstract.Microsoft Visual C++ is a very well suited integrated development environment for professional developers on Win32 platforms. Either MFC library, build-in WinAPI or numerous opportunities for parallel programming are extremely useful in daily work. Authors of the "Visual C++. Recipes for Windows programers." don't focus on describing IDE options, but rather on capabilities which it offers to users. First we learn how to build our own GUI, control the state of the system, then we familiarise with file system, multimedia, register, Windows messages and DLLs. Finally we master selected parallel programming APIs (Threads, OpenMP, Intel Threading Building Blocks and Nvidia CUDA). My contribution is section dedicated to CUDA.
- M. Matuszak, J. Matulewski,
CUDA i czyny,Software Developer's Journal, 01.2010 Abstract.After reading our previous work about CUDA we can write simple kernels. Now we will focus on harder parts of CUDA's programming. We describe basic optimisation techniques, such as using shared memory, coalesced reads from global memory or transferring data from/to graphics card.
- M. Matuszak, J. Matulewski,
Czyn CUDA,Software Developer's Journal, 12.2009, (Cover article) Abstract.In this article we present an introduction to Nvidia's CUDA. CUDA (stands for Compute Unified Device Architecture) is a technology that provides access to computational resources of modern GPUs. We describe architecture of those cards, then a new programming model introduced by Nvidia and finally we show how to write a simple program for CUDA.
Bayesian Networks in Adaptation and Optimization of Behavioral Patterns,May 2013 Abstract.In this thesis, we present several new methods and algorithmic results related to probabilistic graphical models. In the first part, we present a short introduction to graphical models in the context of the thesis results. Our results are summarized and possible further research are pointed out in the last chapter. Finally, we include published papers.
One of the most important result was developed for the strategy optimization in Bayesian influence diagrams. It is a well-known NP-complete problem. The proposed stochastic algorithm generates optimal decision strategies by an iterative self-annealing reinforced search procedure, gradually acquiring new information while driven by information already acquired. At the basis of the method lies the Chen-style stochastic optimization which was originally proposed for travelling salesman problems (TSP). The algorithm, after a substantial extension, is applied to the NP-hard problem of learning Bayesian network structure. Another application of the algorithm is in the NP-hard ramified optimal transport problem.
In Gaussian-network set up, we develop an algorithm for determining optimal transition paths between given configurations of systems consisting of many objects. The method is applied to a system controlling the motion and redeployment between unit's formations and to a realistic transformation between two sequences of character animations in a virtual environment.
Using the framework of polygonal Markov fields, we introduce an image segmentation algorithm. Our algorithm is based on the Markovian optimization dynamics combining the simulated annealing ideas with those of the Chen-style stochastic optimization - in which successive segmentation updates are carried out simultaneously with the adaptive optimization of the local activity functions.
Application of Monte Carlo Methods in Inference Algorithms in Continuous Time Bayesian Networks (in polish),June 2009
Conferences/Workshops [show all]
- GTC 2015, GPU Technology Conference, 16-20.03.2015, San Jose, USA
- ICISIP 2014: The 2nd International Conference on Intelligent Systems and Image Processing 2014, Kitakyushu, Fukuoka, Japan, 26-29.09.2014
- GTC 2014, GPU Technology Conference, 23-27.03.2014, San Jose, USA
- PUMPS 2013, Programming and Tuning Massively Parallel Systems, 8-12.07.2013, Barcelona, Spain
- IwE 2013, Informatics in Education, 5-7.07.2013, Torun, Poland
- WCCE 2013, 10th IFIP World Conference on Computers in Education, 2-5.07.2013, Torun, Poland
- SGSIA, 17th Workshop on Stochastic Geometry, Stereology and Image Analysis, 11-15.06.2013, Torun, Poland
- Tomasz Schreiber's Memorial Session, 10.06.2013, Torun, Poland
- Summer School on Network Science, 20-31.05.2013, Columbia, SC, USA
- FIT 2013, Forum of Theoretical Informatics, 11-14.04.2013, Torun, Poland
- FENS-IBRO-Hertie Winter School: Brain Dynamics and Dynamics of Brain Disease, 09-16.12.2012, Obergurgl, Austria
- Minikonferencja 5, 19-20.10.2011, Wroclaw, Poland
- Vienna Game/AI Conference 2012, 17-19.09.2012, Vienna, Austria
- ICANN 2012, International Conference on Artificial Neural Networks, 11-14.09.2012, Lausanne, Switzerland
- ABS12, Applied Bayesian Statistics School: Stochastic Modelling for Systems Biology, 3-7.09.2012, Pavia, Italy
- ICAISC 2012, 11th International Conference on Artificial Intelligence and Soft Computing, 29.04-3.05.2012, Zakopane, Poland
- Minikonferencja 4, 13.04-14.04.2012, Torun, Poland
- GPUs in Computational Statistics, 25.01.2012, Coventry, United Kingdom
- Minikonferencja 3, 14.10-15.10.2011, Krakow, Poland
- KI 2011, 34th Annual German Conference on Artificial Intelligence, 4.10-7.10.2011, Berlin, Germany
- WGK 2011, I Krajowa Konferencja Wytwarzania Gier Komputerowych, 2.09-4.09.2011, Gdansk, Poland
- TLSM 2011, Toruńska Letnia Szkoła Matematyki, 22.08-26.08.2011, Torun, Poland
- OCNC 2011, Okinawa Computational Neuroscience Course, 13.06-30.06.2011, Okinawa, Japan
- Minikonferencja 2, 8.04-9.04.2011, Poznan, Poland
- KROK W PRZYSZLOSC - stypendia dla doktorantow III edycja, 2-3.12.2010, Torun, Poland
- MASYW'10, Mathematical Models and Methods of Analysis of Concurrent Systems , 19-23.07.2010, Tlen nad Wda, Poland
- NN2010, Summer School on Neural Networks in Classification, Regression and Data Mining, 12-16.07.2010, Porto, Portugal
- ICAISC 2010, 10th International Conference on Artificial Intelligence and Soft Computing, 13-17.06.2010, Zakopane, Poland
- How to model neurons and neural systems? Integrating biophysics, morphology, and connectivity. Second Polish-Norwegian Neuroinformatics Workshop, 14-15.01.2010, Warszawa, Poland
- Komercjalizacja wiedzy (Fundacja Centrum Innowacji FIRE) 17-18.12.2009, Torun, Poland
- KROK W PRZYSZLOSC - stypendia dla doktorantow III edycja, 10.12.2009, Torun, Poland
- Spin-off - biznesowa szansa dla studentow i doktorantow, 26.11.2009, Torun, Poland
- KI 2009, 32nd Annual Conference on Artificial Intelligence, AI and Automation, 15-18.09.2009, Paderborn, Germany
- Tango Conference 2008, First International D language Community Meeting, 26-28.09.2009, Torun, Poland
IwE 2012, Informatyka w Edukacji, 3-4.07.2012, Torun, Poland
- Graphics and Multimedia  (in polish)
- Operating Systems  (in polish)
- Computer Science (High School) [2011-13] (in polish)
- Bartek Zielinski - a Flash developer
- Tomek Stachowiak - a Game developer
- Filip Piekniewski - expert in Neural Networks
- Jarek Piersa - specialist in Neural Networks
- Maja Czokow - PhD student in Computer Science (Spring Systems)