DSpace Repository

Comparing Direct and Indirect Encodings Using Both Raw and Hand-Designed Features in Tetris

Show simple item record

dc.contributor.author Schrum, Jacob
dc.contributor.author Gillespie, Lauren E.
dc.contributor.author Gonzalez, Gabriela R.
dc.date.accessioned 2018-12-10T19:20:40Z
dc.date.available 2018-12-10T19:20:40Z
dc.date.issued 2017-07
dc.identifier.citation Lauren E. Gillespie, Gabriela R. Gonzalez, and Jacob Schrum. 2017. Comparing direct and indirect encodings using both raw and hand-designed features in tetris. In Proceedings of the Genetic and Evolutionary Computation Conference (GECCO '17). ACM, New York, NY, USA, 179-186. DOI: https://doi.org/10.1145/3071178.3071195 en_US
dc.identifier.uri http://hdl.handle.net/11214/219
dc.description.abstract Intelligent agents have a wide range of applications in robotics, video games, and computer simulations. However, fully general agents should function with as little human guidance as possible. Specifically, agents should learn from large collections of raw state variables instead of small collections of hand-designed features. Learning from raw state variables is difficult, but can be easier when agents are aware of the geometry of the input space. Indirect encodings allow agents to take advantage of the geometry of the task, and scale up to large input spaces. This research demonstrates the relative benefits of a direct and indirect encoding using raw or hand-designed features in Tetris, a challenging video game. Specifically, the direct encoding NEAT is compared against the indirect encoding HyperNEAT Both algorithms create neural networks to play the game, but HyperNEAT makes better use of raw screen inputs, due to its ability to generate large networks that take advantage of the domain's geometry. However, hand-designed features lead to higher scores with both algorithms. HyperNEAT makes better use of hand-designed features early in evolution, but NEAT eventually overtakes it. Since each method succeeds in different circumstances, approaches combining the strengths of both should be explored. en_US
dc.language.iso en_US en_US
dc.publisher Proceedings of the Genetic and Evolutionary Computation Conference en_US
dc.subject Computing methodologies en_US
dc.subject Machine learning en_US
dc.subject Machine learning approaches en_US
dc.subject Neural networks en_US
dc.subject Bio-inspired approaches en_US
dc.subject Generative and developmental approaches en_US
dc.title Comparing Direct and Indirect Encodings Using Both Raw and Hand-Designed Features in Tetris en_US
dc.type Article en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search


Browse

My Account

Links