• Ing. Jaroslav Vítků

    Graduated in 2011 in Czech Technical University in Prague, Faculty of Electrical Engineering in Artificial Intelligence. Currently is a PhD student in the CTU, FEE, Department of Cybernetics.

    His research interests include modular agent architectures, hybrid & spiking neural networks, neural engineering, hierarchical problem decomposition, (behavioral and cognitive) robotics, evolutionary algorithms, Artificial Life and generally biologically inspired systems in common. Contact information is here and other information here.

    Current Projects

    - Hybrid Artificial Neural Network Systems Project 
    - Architect – Neuroevolution of Hybrid Artificial Neural Network Systems Project 
    - Nengo+Ros – Simulator of heterogenous systems SW tool

    Finished Projects

    - An Artificial Creature Capable of Learning from Experience in Order to Fulfill More Complex Tasks (Diploma thesis) . Here is new presentation video.
    - Cascade Evolutionary Algorithm (Bachelor thesis) 

    Publications

    2014

    • P. Nahodil and J. Vítků, “Artificial Intelligence and Cognitive Science,” , V. Kvasnička, J. Pospíchal, P. Návrat, D. Chalupa, and L. Clementis, Ed., Faculty of Informatics and Information Technologies SUT in Bratislava, 2014, pp. 159-186, ISBN: 978-80-227-4208-5.
      [Bibtex]
      @INBOOK{oursbookslovakia2014,
        chapter = {Ethology-Inspired Design of Autonomous Agents in Domain of Artificial Life},
        title = {Artificial Intelligence and Cognitive Science},
        publisher = {Faculty of Informatics and Information Technologies SUT in Bratislava},
        year = {2014},
        pages = {159-186},
        editor = {Kvasnička, V. and Pospíchal, J. and Návrat, P. and Chalupa, D. and Clementis, L.},
        author = {Nahodil, P. and Vítků, J.},
        note = {ISBN: 978-80-227-4208-5},
      abstract = {This chapter describes new methods of designing of autonomous agents. We inspire ourselves in fields as Artificial Intelligence, Ethology and Biology, while designing our agents. Typical course of agent’s life is similar to newly born animal, which continuously learns itself: consequently from basic information about its environment towards the ability to solve complex problems. Our latest architecture integrates several learning and action-selection mechanisms into one more complex system. The main advantages of such an agent are in its total autonomy, the ability to gain all information from a surrounding environment. Also, the ability to efficiently decompose potentially huge decision space into a hierarchy of smaller spaces enables the agent to successfully learn and “live” also in very complex domains. Unsupervised learning is triggered mainly by agent’s predefined physiology and intentions which are autonomously created during his life. We present here theoretical background used for creation of our agents, then we mention several main ideas behind our research are presented. Finally, we describe our latest architectures of autonomous agents. Several experiments which were concluded in order to validate the expected abilities of our agents are also presented. One of main contributions of our research is in proposing a new hybrid domain independent hierarchical planner. This planner combines classical planning system with hierarchical reinforcement learning. The ability to accommodate changing ideas about causality allows the creature to exist in and adapt to a dynamic world.}
      }
    • J. Vítků and P. Nahodil, “Automaticky navrhované evoluční architektury chování agentů,” Automa, 2014.
      [Bibtex]
      @article{oursautoma2014,
        author = {Vítků, J. and Nahodil, P.},
        title = {Automaticky navrhované evoluční architektury chování agentů},
        journal = {Automa},
        issue_date = {2014},
        year = {2014},
        numpages = {7},
        publisher = {Automa-časopis pro automatizační techniku, s.r.o.},
        issn = {1210-9592},
        note={Accepted for publication, Author's participation = 70\%}
      }
    • [PDF] [DOI] J. Vítků and P. Nahodil, “Towards Evolutionary Design of Complex Systems Inspired by Nature,” Acta Polytechnica – Journal of Advanced Engineering, vol. 54, iss. 5, pp. 367–377, 2014.
      [Bibtex]
      @article{oursacta2014,
        author = {Vítků, J. and Nahodil, P.},
        title = {Towards Evolutionary Design of Complex Systems Inspired by Nature},
        journal = {Acta Polytechnica - Journal of Advanced Engineering},
        issue_date = {2014},
        year = {2014},
        volume={54},
        number={5},
        pages={367–377},
        numpages = {10},
        publisher = {Czech Technical University in Prague, CTU Publishing House},
        issn = {1805-2363},
        doi={10.14311/AP.2014.54.0367},
        url={https://ojs.cvut.cz/ojs/index.php/ap/article/view/AP.2014.54.0367},
        ISSN={1210-2709},
        note={Author's participation = 70\%}
      }
    • P. Nahodil and J. Vítků, “Evoluční architektura chování umělých bytosti – agentů v daném prostředí,” in Kognitivní věda a umělý život XIV (KUZ XIV), Zaječí u Břeclavi: Slezská univerzita v Opavě, 2014, pp. 155–164, Author’s participation = 40\%.
      [Bibtex]
      @INPROCEEDINGS{ourskuz2014,
        author = {Nahodil, P. and Vítků, J.},
        title = {Evoluční architektura chování umělých bytosti - agentů v daném prostředí},
        booktitle = {Kognitivní věda a umělý život XIV (KUZ XIV)},
        year = {2014},
        pages = {155–164},
        address = {Zaječí u Břeclavi: Slezská univerzita v Opavě},
        abstract = {Podobně jako jsou neuronové sítě složeny z neuronů, naše architektury autonomních agentů jsou složené z „neurálních modulů“, které jsou mezi sebou propojeny stejně jako neurony v umělých neuronových sítích. Tyto hybridní sítě je pak možné optimalizovat pomocí neuro-evolučních metod a tím navrhovat nové řídící architektury. Zde se zaměřujeme na architekturu agenta obsahující modul posilovaného učení a zdroj motivace. Modul posilovaného učení generuje akce a jejich konsekvence se učí zodměn přijatých od ostatních modulů (fyziologie agenta). Je zde popsán automatický návrh nových architektur pomocí algoritmu inspirovaného neuro-evolucí. Princip funkce dvou takto automaticky navržených architektur je porovnán s ručně zapojenou architekturou agenta.},
      isbn={978-80-7248-951-0},
      note={Author's participation = 40\%}
      }
    • [PDF] [DOI] J. Vítků and P. Nahodil, “Reusable Reinforcement Learning for Modular Self Motivated Agents,” in Proceedings of 28th European Conference on Modeling and Simulation, Brescia, Italy, 2014, pp. 352-358, Author’s participation = 60\%, Note: will be indexed in WoS as previous issues.
      [Bibtex]
      @INPROCEEDINGS{oursecms2014,
        author = {Vítků, J. and Nahodil, P.},
        title = {Reusable Reinforcement Learning for Modular Self Motivated Agents},
        booktitle = {Proceedings of 28th European Conference on Modeling and Simulation},
        year = {2014},
        editor = {Flaminio Squazzoni and Fabio Baronio and Claudia Archetti and Marco Castellani},
        pages = {352-358},
        address = {Brescia, Italy},
        publisher = {European Council for Modelling and Simulation},
        abstract = {Presented topic is from the research fields called Artificial Life and Artificial Intelligence (AI). In this paper, there is presented novel approach to designing agent architectures with its requirements. The approach in inspired by inherited modularity of biological brains and agent architectures are represented here as set of given reusable modules connected into a particular topology. This paper presents design of two particular modules for future use in more complex architectures. The modules are used for implementing model-free motivation-driven Reinforcement Learning (RL). First, the novel framework for these architectures is described together with a used simulator. Then, the design of two new reusable domain-independent components of agent architectures is described. Finally, experimental validation of these new components and their future use is mentioned.},
        doi = {10.7148/2014-0352},
        url = {http://dx.doi.org/10.7148/2014-0352},
        isbn={978-0-9564944-8-1},
        note={Author's participation = 60\%, Note: will be indexed in WoS as previous issues},
        keywords = {WoS, Scopus},
      }
    • [DOI] J. Vítků and P. Nahodil, “Q-Learning Algorithm Module in Hybrid Artificial Neural Network Systems,” in Modern Trends and Techniques in Computer Science, 2014, pp. 117-127, Author’s participation = 60\%, Note: will be added in the WoS and Scopus as other proceedings in this Springer series (\urlhttp://www.springer.com/series/11156)..
      [Bibtex]
      @INPROCEEDINGS{ourscsoc2014,
        author = {Vítků, J. and Nahodil, P.},
        title={Q-Learning Algorithm Module in Hybrid Artificial Neural Network Systems},
        booktitle={Modern Trends and Techniques in Computer Science},
        year = {2014},
        series={Springer, Advances in Intelligent Systems and Computing},
        volume={285},
        editor={Silhavy, Radek and Senkerik, Roman and Oplatkova, Zuzana Kominkova and Silhavy, Petr and Prokopova, Zdenka},
        pages = {117-127},
        publisher={Springer International Publishing},
        abstract = {Presented topic is from the research field called Artificial Life, but contributes also to the field of Artificial Intelligence (AI), Robotics and potentially into many other aspects of research. In this paper, there is reviewed and tested new approach to autonomous design of agent architectures. This novel approach is inspired by inherited modularity of biological brains. During designing of new brains, the evolution is not directly connecting individual neurons. Rather than that, it composes new brains by connecting larger, widely reused areas (modules). In this approach, agent architectures are represented as hybrid artificial neural networks composed of heterogeneous modules. Each module can implement different selected algorithm. Rather than describing this framework, this paper focuses on designing of one module. Such a module represents one component of hybrid neural network and can seamlessly integrate a selected algorithm into the node. The course of design of such a module is described on example of discrete reinforcement learning algorithm. The requirements posed by the framework are presented, the modifications on the classical version of algorithm are mentioned and then the resulting performance of module with expectations is evaluated. Finally, the future use cases of this module are described.},
        doi={10.1007/978-3-319-06740-7_11},
        url={http://dx.doi.org/10.1007/978-3-319-06740-7_11},
        keywords={Agent; Architecture; Artificial life; Creature; Behaviour; Hybrid; Neural networks; Evolution, WoS, Scopus},
        isbn = {978-3-319-06739-1},
        note={Author's participation = 60\%, Note: will be added in the WoS and Scopus as other proceedings in this Springer series (\url{http://www.springer.com/series/11156}).}
      }

    2013

    • [PDF] [DOI] J. Vítků and P. Nahodil, “Autonomous Design of Modular Intelligent Systems,” in 27th European Conference on Modelling and Simulation ECMS 2013, Alesund, 2013, pp. 379-389, Author’s participation = 60\%.
      [Bibtex]
      @INPROCEEDINGS{oursecms2013,
        author = {Vítků, J. and Nahodil, P.},
        title = {Autonomous Design of Modular Intelligent Systems},
        booktitle = {27th European Conference on Modelling and Simulation ECMS 2013},
        year = {2013},
        editor = {Webjørn Rekdalsbakken and Robin T. Bye and Houxiang Zhang},
        pages = {379-389},
        address = {Alesund},
        publisher = {European Council for Modelling and Simulation},
        abstract = {We propose our original system capable of autonomous design of general-purpose
        complex modular hybrid systems. The resulting hybrid systems will
        be able to employ various techniques of learning, decision-making,
        prediction etc. Presented topic is from Artificial Life domain, but
        contributes also to fields such as Artificial Intelligence, Biology,
        Computational Neuroscience, Ethology, Cybernetics and potentially
        into many other aspects of research. The autonomous design is implemented
        as an optimization of system topology with respect to given problem.
        The principle of design is based on modified neuro-evolution and
        can be compared to modular neural networks. One of the main requirements
        is standardization of communication between very different subsystems.
        Here, each subsystem - module implements arbitrary algorithm and
        is treated as a Multiple-Input Multiple-Output subsystem. First,
        the design of simulator used is described, then the basic principle
        of hybrid networks is explained with it benefits and drawbacks. Finally,
        simple example is mentioned.},
          doi={10.7148/2013-0379},
          url={http://dx.doi.org/10.7148/2013-0379},
          isbn={978-0-9564944-6-7},
         note={Author's participation = 60\%},
        keywords={WoS, Scopus}
      }
    • P. Nahodil and J. Vítků, “Hybridní neuronové systémy pro návrh architektur autonomních agentů v oblasti umělého života,” in Kognícia a Umelý Život XIII, Opava, Stará Lesná, 2013, pp. 197-204, Author’s participation = 40\%.
      [Bibtex]
      @INPROCEEDINGS{ourskuz2013,
        author = {Nahodil, P. and Vítků, J.},
        title = {Hybridní neuronové systémy pro návrh architektur autonomních agentů
        v oblasti umělého života},
        booktitle = {Kognícia a Umelý Život XIII},
        year = {2013},
        pages = {197-204},
        address = {Opava, Stará Lesná},
        abstract = {V současné době existuje mnoho typů různých algoritmů a jejich implementací.
        Za účelem zvýšení efektivity výzkumu v oblasti kompozice modulárních
        inteligentních systémů se snažíme sjednotit komunikaci mezi různými
        algoritmy. Algoritmus je zde viděn jako subsystém a komunikace mezi
        subsystémy je definována podobně jako komunikace v umělých neuronových
        sítích. Libovolnou kombinací těchto subsystémů (modulů) jsme poté
        schopni vytvářet nové modulární systémy. Za účelem praktického ověření
        chování výsledných architektur jsme vytvořili nový simulátor kombinací
        dvou stávajících softwarových nástrojů. Práce je z oblasti umělého
        života, proto se zde zabýváme především možnostmi kompozice architektur
        autonomních agentů. Simulovat je však možné modulární systémy napříč
        vědními obory. Definování těchto modulárních podobně jako neuronové
        sítě má následující hlavní důsledky: Přístup nám umožňuje libovolně
        kombinovat způsoby návrhu architektur „zdola nahoru“ a „shora dolů“.
        Výsledná síť má poté mnohem menší komplexitu topologie oproti klasické
        umělé neuronové síti. Navíc, jednotná komunikace nám umožňuje zkoumat
        možnosti automatického návrhu těchto modulárních systémů.},
        owner = {j},
        timestamp = {2013.07.01},
      isbn={978-80-7248-863-6},
      note={Author's participation = 40\%}
      }

    2012

    • [PDF] [DOI] P. Nahodil and J. Vítků, “Learning of Autonomous Agent in Virtual Environment,” in 26th European Conference on Modelling and Simulation (ECMS), 2012, pp. 373-379, Author’s participation = 40\%.
      [Bibtex]
      @INPROCEEDINGS{oursecms2012,
        author = {Nahodil, P. and Vítků, J.},
        title = {Learning of Autonomous Agent in Virtual Environment},
        booktitle = {26th European Conference on Modelling and Simulation (ECMS)},
        year = {2012},
        pages = {373-379},
        abstract = {Presented topic is from area of development of artificial creatures
        and proposes new architecture of autonomous agent. The work builds
        on a research of the latest approaches to Artificial Life, realized
        by the Department of Cybernetics, CTU in Prague in the last twenty
        years. This architecture design combines knowledge from Artificial
        Intelligence (AI), Ethology, Artificial Life (ALife) and Intelligent
        Robotics. From the field of classical AI, the fusion of reinforcement
        learning, planning and artificial neural network into one more complex
        control system was used here. The main principle of its function
        is inspired by the field of Ethology, this means that life of given
        agent tries to be similar to life of an animal in the Nature, where
        animal learns relatively autonomously from simpler principles towards
        the more complex ones. The architecture supports on-line learning
        of all knowledge from the scratch, while the core principle is in
        hierarchical Reinforcement Learning (RL), this action hierarchy is
        created autonomously based solely on agents interaction with an environment.
        The main key idea behind this approach is in original implementation
        of a domain independent hierarchical planner. Our planner is able
        to operate with behaviors learned by the RL. It means that an autonomously
        gained hierarchy of actions can be used not only by action selection
        mechanisms based on the reinforcement learning, but also by a planning
        system. This gives the agent ability to utilize high-level deliberative
        problem solving based solely on his experiences. In order to deal
        with higher-level control rather than a sensory system, the life
        of agent was simulated in a virtual environment.},
        doi = {10.7148/2012-0373-0379},
        owner = {j},
        timestamp = {2012.12.09},
        url={http://www.scs-europe.net/conf/ecms2012/ecms2012%20accepted%20papers/is_ECMS_0084.pdf},
        keywords = {WoS, Scopus},
        isbn={978-0-9564944-4-3},
        note={Author's participation = 40\%}
      }
    • [PDF] J. Vítků, “Ethology-Inspired Advanced Problem Solving Mechanism,” in POSTER 2012 -16th International Student Conference on Electrical Engineering, 2012, pp. 1-6, ISBN 978-80-01-05043-9.
      [Bibtex]
      @INPROCEEDINGS{jvitkuposter2012,
        author = {Vítků, J.},
        title = {Ethology-Inspired Advanced Problem Solving Mechanism},
        booktitle = {POSTER 2012 -16th International Student Conference on Electrical
        Engineering},
        year = {2012},
        pages = {1-6},
        organization = {Czech Technical University in Prague},
        note = {ISBN 978-80-01-05043-9},
        abstract = {Presented topic is from the research field called Artificial Life
        (ALife), but contributes also to the field of Artificial Intelligence
        (AI). We have built a new architecture of autonomous agent. Our agent
        is capable of fully autonomous learning. The course of learning is
        similar to newly born animal. The agent learns everything itself,
        from basic tasks towards the more complex knowledge. This newly gained
        knowledge is stored in the hierarchy of abstract actions. The main
        innovation in our approach can be seen in our new way how to implement
        hybrid hierarchical planner. This planner combines Reinforcement
        Learning (RL) and classical planning. The resulting hierarchical
        planner is domain independent, because it is capable to adapt itself
        to the given domain. We have tested this on the artificial agent
        within the virtual simulation environment.},
        owner = {j},
        timestamp = {2012.12.09}
      }
    • J. Vítků and P. Nahodil, “Nové hybridní rozhodovací mechanismy v oblasti umělého života,” in Kognice a umělý život (KUZ XII), Průhonice, 2012, pp. 254-263, Author’s participation = 60\%.
      [Bibtex]
      @INPROCEEDINGS{ourskuz2012,
        author = {Vítků, J. and Nahodil, P.},
        title = {Nové hybridní rozhodovací mechanismy v oblasti umělého života},
        booktitle = {Kognice a umělý život (KUZ XII)},
        year = {2012},
        pages = {254-263},
        address = {Průhonice},
        publisher = {Agentura Action M},
        abstract = {V tomto článku chceme představit naše nové způsoby návrhu autonomních
        agentů. Námi navrhované architektury jsou inspirovány především umělou
        inteligencí, biologií a etologií, kde průběh života agenta je podobný
        životu zvířete v přírodě, které se samo učí postupně od jednoduchých
        znalostí směrem ke schopnostem plnění složitějších úloh. Věříme,
        že budoucnost vývoje je v hybridním přístupu, a proto naše architektury
        integrují několik typů učících algoritmů a mechanismů pro selekci
        akcí do jednoho systému. Prezentujeme zde především tři na sebe navazující
        výsledky našeho aktuálního výzkumu. V prvních dvou případech se jedná
        o nové architektury autonomních agentů, v posledním případě jde o
        současný výzkum v oblasti hybridních kognitivních architektur.},
        owner = {j},
        timestamp = {2012.12.09},
        isbn={978-80-86742-34-2},
        note={Author's participation = 60\%}
      }
    • [DOI] P. Nahodil and J. Vítků, “How to Design an Autonomous Creature Based on Original Artificial Life Approaches,” in Beyond Artificial Intelligence, Plzeň: Springer, 2012, pp. 161-180, Author’s participation = 40\%.
      [Bibtex]
      @INCOLLECTION{oursbaispringer2012,
        author = {Nahodil, P. and Vítků, J.},
        title = {How to Design an Autonomous Creature Based on Original Artificial
        Life Approaches},
        booktitle = {Beyond Artificial Intelligence},
        publisher = {Springer},
        year = {2012},
        pages = {161-180},
        address = {Plzeň},
        abstract = {We introduce new approaches for creating of autonomous agents. The
        life of such creature is very similar to the animal’s life in the
        Nature, which learns autonomously from the simple tasks towards the
        more complex ones and is inspired in AI, Biology and Ethology. We
        present our established design of artificial creature, capable of
        learning from its experience in order to fulfill more complex tasks,
        which is based mainly on ethology. It integrates several types of
        action-selection mechanisms and learning into one system. The main
        advantages of the architecture is its autonomy, the ability to gain
        all information from the environment and decomposition of the decision
        space into the hierarchy of abstract actions, which dramatically
        reduces the total size of decision space. The agent learns how to
        exploit the environment continuously, where the learning of new abilities
        is driven by his physiology, autonomously created intentions, planner
        and neural network.},
        doi = {10.1007/978-3-642-34422-0_11},
        owner = {j},
        timestamp = {2012.12.09},
        url = {http://link.springer.com/content/pdf/10.1007/978-3-642-34422-0_11.pdf},
        keywords = {WoS},
        issn = {2193-9411},
      isbn={978-3-642-34421-3},
      note={Author's participation = 40\%}
      }
    • [DOI] P. Nahodil and J. Vítků, “Novel Theory and Simulations of Anticipatory Behaviour in Artificial Life Domain,” in Advances in Intelligent Modelling and Simulation, Springer, 2012, pp. 131-164, Author’s participation = 35\%.
      [Bibtex]
      @INCOLLECTION{oursspringer2012,
        author = {Nahodil, P. and Vítků, J.},
        title = {Novel Theory and Simulations of Anticipatory Behaviour in Artificial Life Domain},
        booktitle = {Advances in Intelligent Modelling and Simulation},
        publisher = {Springer},
        year = {2012},
        pages = {131-164},
        abstract = {Recently, anticipation and anticipatory learning systems have gained
        increasing attention in the field of artificial intelligence. Anticipation
        observed in animals combined with multi-agent systems and artificial
        life gave birth to the anticipatory behaviour. This is a broad multidisciplinary
        topic. In this work, we will first introduce the topic of anticipation
        and will describe which scientific field it belongs to. The state
        of the art on the field of anticipation in details and mention works
        and theories that contributed to our approach will be described.
        The parts important for presented research are further detailed probed
        in terms of algorithms and mechanisms. Designed multi-level anticipatory
        behaviour approach is based on the current understanding of anticipation
        from both the artificial intelligence and the biology point of view.
        Original thought is that we have to use not one but multiple levels
        of unconscious and conscious anticipation in a creature design. The
        aim of this chapter is not only to extensively present all the achieved
        results but also to demonstrate the thinking behind. Primary industrial
        applications of this 8-factor anticipation framework design are intelligent
        robotics and smart grids.},
        doi = {10.1007/978-3-642-28888-3_6},
        owner = {j},
        timestamp = {2012.12.09},
        url = {http://link.springer.com/chapter/10.1007%2F978-3-642-28888-3_6#page-1},
        keywords={Scopus,WoS,nophdrelated},
        isbn = {978-3-642-28887-6},
      note={Author's participation = 35\%}
      }

    2011

    • [View externally] J. Vítků, “An Artificial Creature Capable of Learning from Experience in Order to Fulfill More Complex Tasks,” Diploma Thesis, Czech Technical University in Prague, Faculty of Electrical Engineering, Dept. of Cybernetics, Supervisor: Doc. Ing. Nahodil Pavel CSc. (in English), 2011.
      [Bibtex]
      @MASTERSTHESIS{jvitkudt2011,
        author = {Vítků, J.},
        title = {An Artificial Creature Capable of Learning from Experience in Order
        to Fulfill More Complex Tasks},
        school = {Czech Technical University in Prague, Faculty of Electrical Engineering,
        Dept. of Cybernetics},
        year = {2011},
        type = {Diploma Thesis},
        note = {Supervisor: Doc. Ing. Nahodil Pavel CSc. (in English)},
        abstract = {Presented Diploma Thesis is from area of development of artificial
        creatures. The work builds on a research of the latest approaches
        to artificial life, realized by the Department of Cybernetics, CTU
        in Prague, under the leadership of Pavel Nahodil in the last twenty
        years.
        
        The main feature of this architecture should be it’s total autonomy,
        ability to gain all information from the surrounding environment
        and effective information filtering and classification. Agent is
        supposed to operate based only on the sensory input and by it’s actuator
        system, because of the properties meant above the resulting architecture
        should be almost fully independent on the concrete area and form
        of use. Consequently it is unimportant whether the agent is embodied
        in some robotic system, intelligent house, or just operates in some
        virtual environment. Thanks to the fact that all the designer has
        to specify is just the sensory layer, actuator layer, and agent’s
        needs, the architecture should be convenient especially in unknown
        environments, where some more complex task has to be fulfilled.
        
        This agent architecture is based on the layered model, combining various
        approaches on different layers. The life of agent is similar to a
        newly born animal, which explores new and unknown environment, learns
        from experiences and links the newly learned abilities to it’s needs
        in order to survive and increase effectiveness of it’s behavior.
        New knowledge is learned simultaneously on various levels of abstraction
        using different learning approaches.
        
        Besides the proposing the agent architecture, I hope that this thesis
        should be con- tribution also in the approach of combining some various
        types of decision making, the resulting system exploits their benefits
        and inhibits their weaknesses. One of the main features is an alternative
        implementation of system similar to reactive and hierarchical planning.
        The system combines hierarchical reinforcement learning with planning
        engine. The main benefit should be the domain independency of this
        planner, because the entire hierarchy of actions is created completely
        autonomously, the other benefit is tight connec- tion of these two
        systems, so the boundary how big part of given task should be created
        deliberatively and which rather reactively (based on the concrete
        actual situation), can be chosen online based on the specific agent’s
        preferences. That’s why I hope that the proposed part of architecture
        could be used also as an advanced domain independent hierarchical
        planning engine with autonomous learning from the experiences.},
        owner = {j},
        timestamp = {2012.01.26},
        url = {http://cyber.felk.cvut.cz/research/theses/detail.phtml?id=176},
        keywords = {nophdrelated}
      }
    • [PDF] [View externally] P. Nahodil and J. Vítků, “New Way How to Build an Autonomous Creatures,” in International Conference: Beyond Artificial Intelligence, Interdisciplinary Aspects of Artificial Intelligence, Pilsen, 2011, pp. 34-41, Author’s participation = 40\%.
      [Bibtex]
      @INPROCEEDINGS{oursbai2011,
        author = {Nahodil, P. and Vítků, J.},
        title = {New Way How to Build an Autonomous Creatures},
        booktitle = {International Conference: Beyond Artificial Intelligence, Interdisciplinary
        Aspects of Artificial Intelligence},
        year = {2011},
        pages = {34-41},
        address = {Pilsen},
        month = {December},
        abstract = {We introduce new approaches for creating of autonomous agents. The
        life of agent created by us is very similar to the animal’s life
        in the Nature, which learns autonomously from the simple tasks towards
        the more complex ones and is inspired in AI, Biology and Ethology.
        We present our proved design of artificial creature, capable of learning
        from experience in order to fulfil more complex tasks, which is based
        mainly on ethology. It integrates several types of action selection
        mechanisms and learning into one system. The main advantages are
        in its autonomy, ability to gain all information from the environment
        and decomposition of the decision space into the hierarchy of abstract
        actions, which dramatically reduces the total size of decision space.
        The agent learns how to exploit the environment continuously, where
        the learning of new abilities is driven by his physiology, autonomously
        created intentions, planner and neural network.},
        owner = {j},
        timestamp = {2012.12.09},
        url = {http://www.kky.zcu.cz/en/publications/1/JanRomportl_2011_BeyondAI.pdf},
      note={Author's participation = 40\%}
      }

    2009

    • [View externally] J. Vítků, “Cascade Evolutionary Algoritm,” Bachelor Thesis, Czech Technical University in Prague, Faculty of Electrical Engineering, dept. of Cybernetics, Supervisor: Ing. Jiří Kubalík, Ph.D. (in English), 2009.
      [Bibtex]
      @MASTERSTHESIS{jvitkubt2009,
        author = {Jaroslav Vítků},
        title = {Cascade Evolutionary Algoritm},
        school = {Czech Technical University in Prague, Faculty of Electrical Engineering,
        dept. of Cybernetics},
        year = {2009},
        type = {Bachelor Thesis},
        note = {Supervisor: Ing. Jiří Kubalík, Ph.D. (in English)},
        abstract = {The aim of this thesis is to propose and review Cascade Evolutionary
        Algorithm, special case of Parallel Evolutionary Algorithms, which
        should prevent the premature convergence of population better than
        the currently used topologies. This algorithm was designed, implemented
        and experimentally configured, after that some tests on various static
        and dynamic problems were made. The results were compared with one
        of the common Parallel Evolutionary Algorithm topologies.},
        owner = {j},
        timestamp = {2012.01.26},
        url = {http://cyber.felk.cvut.cz/research/theses/detail.phtml?id=85},
        keywords = {nophdrelated}
      }