Evolving Game-Specific UCB Alternatives for General Video Game Playing

Created by W.Langdon from gp-bibliography.bib Revision:1.3973

@InProceedings{Bravi:2017:evoApplications,
  author =       "Ivan Bravi and Ahmed Khalifa and 
                 Christoffer Holmgard and Julian Togelius",
  title =        "Evolving Game-Specific UCB Alternatives for General
                 Video Game Playing",
  booktitle =    "20th European Conference on the Applications of
                 Evolutionary Computation",
  year =         "2017",
  editor =       "Giovanni Squillero",
  series =       "LNCS",
  volume =       "10199",
  publisher =    "Springer",
  pages =        "393--406",
  address =      "Amsterdam",
  month =        "19-21 " # apr,
  organisation = "Species",
  keywords =     "genetic algorithms, genetic programming, General AI,
                 MTCS, Monte-Carlo Tree Search",
  DOI =          "doi:10.1007/978-3-319-55849-3_26",
  abstract =     "At the core of the most popular version of the Monte
                 Carlo Tree Search (MCTS) algorithm is the UCB1 (Upper
                 Confidence Bound) equation. This equation decides which
                 node to explore next, and therefore shapes the
                 behaviour of the search process. If the UCB1 equation
                 is replaced with another equation, the behavior of the
                 MCTS algorithm changes, which might increase its
                 performance on certain problems (and decrease it on
                 others). In this paper, we use genetic programming to
                 evolve replacements to the UCB1 equation targeted at
                 playing individual games in the General Video Game AI
                 (GVGAI) Framework. Each equation is evolved to maximize
                 playing strength in a single game, but is then also
                 tested on all other games in our test set. For every
                 game included in the experiments, we found a UCB
                 replacement that performs significantly better than
                 standard UCB1. Additionally, evolved UCB replacements
                 also tend to improve performance in some GVGAI games
                 for which they are not evolved, showing that
                 improvements generalize across games to clusters of
                 games with similar game mechanics or algorithmic
                 performance. Such an evolved portfolio of UCB
                 variations could be useful for a hyper-heuristic
                 game-playing agent, allowing it to select the most
                 appropriate heuristics for particular games or problems
                 in general.",
  notes =        "Dipartimento di Elettronica, Informatica e
                 BioingegneriaPolitecnico di
                 MilanoMilanoItaly

                 EvoApplications2017 held in conjunction with
                 EuroGP'2017, EvoCOP2017 and EvoMusArt2017
                 http://www.evostar.org/2017/cfp_evoapps.php.",
}

Genetic Programming entries for Ivan Bravi Ahmed Abdel Samea Khalifa Christoffer Holmgard Julian Togelius

Citations