~~NOTOC~~ ====== CI in Games ======
Solving Problems in Games
by Means of
Computational Intelligence
===== Research Context ===== * Deep Reinforcement Learning * Procedural Content Generation * Generative Models * Deep Learning ===== Contacts ===== * [[staff:rudolph|Prof. Dr. Günter Rudolph]] * [[staff:pleines|Marco Pleines]] * [[staff:nicolas_fischoeder|Nicolas Fischöder]] ===== Publications ===== === Conference Articles (peer reviewed) === == 2020 == * Marco Pleines, Jenia Jitsev, Mike Preuss, Frank Zimmer. [[https://arxiv.org/abs/2004.00567|Obstacle Tower Without Human Demonstrations: How Far a Deep Feed-Forward Network Goes with Reinforcement Learning]]. In CoG 2020 Proceedings, IEEE. Best Paper Candidate. == 2019 == * Marco Pleines, Frank Zimmer, Vincent-Pierre Berges. [[http://www.ieee-cog.org/papers/paper_22.pdf|Action Spaces in Deep Reinforcement Learning to Mimic Human Input Devices]]. In CoG 2019 Proceedings, IEEE. === Competitions === == 2019 == * Marco Pleines, Mike Preuss, Jenia Jitsev, Frank Zimmer, Jonathan Indetzki. [[https://youtu.be/P2rBDHBHxcM|Rising to the Obstacle Tower Challenge]]. In CoG 2019 Short Video Competition, IEEE ===== Theses (Abschlussarbeiten) ===== Anybody who is interested in writing her or his thesis under the guidance of the above listed contacts has to submit a well written (not perfect) exposé. The exposé has to be made up of the following items: * 2-3 DIN A4 pages * Content * Context of the thesis * Goal and scope of the thesis * Problems and challenges to be faced * Relevance of conducting the thesis * Initial approach * Preliminary outline of the thesis * Preliminary list of literature Depending on the current interest rate, submitted exposés can be improved for a few iterations until it is good enough for the proposed thesis to be accepted and advised by us. __Note that we cannot guarantee a spot under our guidance beforehand__, because the demand of writing a thesis at our chair is usually much greater than our capacity. === Bachelor Theses === == 2022 == * Leon Swazinna. Evaluation of the MA-POCA Algorithm in a Competitive Reinforcement Learning Environment.\\ Advisors: Rudolph, Pleines. == 2021 == * Alisa Gromova: Training Multiple Agents in a Soccer Environment using Deep Reinforcement Learning and Self-Play.\\ Advisors: Rudolph, Pleines. * Markus Grigull: Sim-to-Real Transfer eines Reinforcement Learning Ansatzes zur mechanischen Steuerung eines Gamepads.\\ Advisors: Rudolph, Pleines. == 2020 == * Matthias Pallasch. Curiosity-driven Exploration mit Reinforcement Learning in einer CoinRun Umwelt.\\ Advisors: Rudolph, Pleines. * Vanessa Speeth. Entwicklung eines Agenten für das Spiel Azul basierend auf dem Advanced-Actor-Critc Ansatz.\\ Advisors: Rudolph, Pleines. * Wentao Li. Applying Curriculum and Reinforcement Learning to a Marble Labyrinth Environment.\\ Advisors: Rudolph, Pleines. == 2019 == * Till Musshoff. Vergleich der Lersperformanz von Proximal Policy Optimization und Behavioral Cloning.\\ Advisors: Rudolph, Pleines. * Marius Brinkmann. Evaluation der Reinforcement Learning-Algorithmen DQN und PPO in einer Ballwurf-Umwelt.\\ Advisors: Rudolph, Pleines. === Master Theses === == 2022 == * Marcel Schyma. Kontextunabhängige prozedurale Szenen- und Inhaltsgenerierung.\\ Advisors: Rudolph, Pleines. == 2021 == * Jonas Schumacher: Deep Reinforcement Learning für Stichspiele mit imperfekter Information / Deep Reinforcement Learning for Trick-Taking Games with Imperfect Information.\\ Advisors: Rudolph, Pleines. ===== Teaching ===== ==== Fachprojekt (technical project) Digital Entertainment Technologies ==== * [[https://ls11-www.cs.tu-dortmund.de/de/rudolph/lehre/fp_det_ws_21_22|WiSe 2021/2022]] * Teacher: Patrick Dinklage * [[https://ls11-www.cs.tu-dortmund.de/de/rudolph/lehre/fp_det_ss_21|SoSe 2021]] * Teacher: Patrick Dinklage * [[https://ls11-www.cs.tu-dortmund.de/de/rudolph/lehre/fp_det_ws20_21|WiSe 2020/2021]] * Teacher: Marco Pleines * [[https://ls11-www.cs.tu-dortmund.de/de/rudolph/lehre/fp_det_ss20|SoSe 2020]] * Teacher: Marco Pleines * [[https://ls11-www.cs.tu-dortmund.de/de/rudolph/lehre/fp_det_ws19_20|WiSe 2019/2020]] * Teacher: Marco Pleines * [[https://ls11-www.cs.tu-dortmund.de/de/rudolph/lehre/fp_det_ss19|SoSe 2019]] * Teacher: Marco Pleines ==== Project Groups ==== === PG 642: Verteiltes Deep Reinforcement Learning System zum Trainieren von Game AI === The goal of this project group is to train agents to play Rocket League using a distributed Deep Reinforcement Learning system. Training directly on Rocket League comes with many issues. Therefore, the game is reimplemented in Unity. This raises the challenge of transferring the learned behavior in Unity to Rocket League, which is called a sim-to-sim transfer. As this project is still ongoing, there are no outcomes to be presented yet. ===== Links ===== ==== Games for Computational Intelligence Research ==== * [[https://github.com/Baekalfen/PyBoy|PyBoy]] * [[https://dockhorn.antares.uberspace.de/wordpress/|Blizzard Hearthstone]] [[https://hearthsim.info/|HearthSim]] * [[https://github.com/Unity-Technologies/ml-agents|Unity ML-Agents]] * [[https://github.com/Unity-Technologies/obstacle-tower-env|Unity Obstacle Tower]] * [[https://github.com/beyretb/AnimalAI-Olympics|AnimalAI-Olympics]] * [[https://gym.openai.com/|OpenAI Gym]] * [[https://retro.readthedocs.io/en/latest/index.html|Gym Retro]] * [[https://openai.com/blog/quantifying-generalization-in-reinforcement-learning/|CoinRun]] * [[https://bwapi.github.io/|Blizzard StarCraft: Brood War]] * [[http://wiki.sc2ai.net/Main_Page|Blizzard StarCraft 2]] * [[https://developer.valvesoftware.com/wiki/Dota_Bot_Scripting|Valve DotA 2]] * [[https://aibirds.org/basic-game-playing-software.html|Angry Birds!]] * [[http://vizdoom.cs.put.edu.pl/|DOOM]] * [[http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.103.3645|Quake 3]] * [[https://github.com/Microsoft/malmo|Microsoft Minecraft]] * [[https://blog.openai.com/retro-contest/|Sonic]] * [[http://www.rlbot.org/|Rocket League]] * [[https://pyplanet.readthedocs.io/en/latest/|Trackmania Nations Forever]] * [[https://ai.googleblog.com/2019/06/introducing-google-research-football.html|Google Research Football]] * [[http://modiki.civfanatics.com/index.php?title=Civ4_Python|Civilization IV]] * [[https://pokemonshowdown.com/|Pokemon]] * [[http://www.ice.ci.ritsumei.ac.jp/~ftgaic/|FightingIce]] * [[http://torcs.sourceforge.net/|The Open Racing Car Simulator]] * [[http://www.gvgai.net/|General Video Game AI]] * [[https://osgameclones.com/|Open Source Game Clones]] * [[https://github.com/elvisyjlin/lf2gym|Little Fighter 2]] * [[https://github.com/vladfi1/gym-dolphin|GameCube and Wii Games]] * [[https://github.com/Nordeus/heroic-rl|Heroic Magic Duel]] * [[https://github.com/MiscellaneousStuff/pylol|League of Legends]] * [[https://pypi.org/project/catanatron/|Catan]] * [[https://sites.google.com/view/arena-unity/home/learning-environments|Arena]] * [[https://github.com/LucasAlegre/sumo-rl|Traffic Control]] * [[https://github.com/apigott/CityLearn|CityLearn]]