Language Understanding for Text-based Games Using Deep Reinforcement Learning. Narasimhan, K., Kulkarni, T., & Barzilay, R. arXiv.org, cs.CL, 2015. abstract bibtex In this paper, we consider the task of learning control policies for text-based games. In these games, all interactions in the virtual world are through text and the underlying state is not observed. The resulting language barrier makes such environments challenging for automatic game players. We employ a deep reinforcement learning framework to jointly learn state representations and action policies using game rewards as feedback. This framework enables us to map text descriptions into vector representations that capture the semantics of the game states. We evaluate our approach on two game worlds, comparing against a baseline with a bag-of-words state representation. Our algorithm outperforms the baseline on quest completion by 54% on a newly created world and by 14% on a pre-existing fantasy game.
@Article{Narasimhan2015,
author = {Narasimhan, Karthik and Kulkarni, Tejas and Barzilay, Regina},
title = {Language Understanding for Text-based Games Using Deep Reinforcement Learning},
journal = {arXiv.org},
volume = {cs.CL},
number = {},
pages = {},
year = {2015},
abstract = {In this paper, we consider the task of learning control policies for text-based games. In these games, all interactions in the virtual world are through text and the underlying state is not observed. The resulting language barrier makes such environments challenging for automatic game players. We employ a deep reinforcement learning framework to jointly learn state representations and action policies using game rewards as feedback. This framework enables us to map text descriptions into vector representations that capture the semantics of the game states. We evaluate our approach on two game worlds, comparing against a baseline with a bag-of-words state representation. Our algorithm outperforms the baseline on quest completion by 54\% on a newly created world and by 14\% on a pre-existing fantasy game.},
location = {},
keywords = {}}
Downloads: 0
{"_id":"X8NzTWyRocpsSCQLw","bibbaseid":"narasimhan-kulkarni-barzilay-languageunderstandingfortextbasedgamesusingdeepreinforcementlearning-2015","authorIDs":[],"author_short":["Narasimhan, K.","Kulkarni, T.","Barzilay, R."],"bibdata":{"bibtype":"article","type":"article","author":[{"propositions":[],"lastnames":["Narasimhan"],"firstnames":["Karthik"],"suffixes":[]},{"propositions":[],"lastnames":["Kulkarni"],"firstnames":["Tejas"],"suffixes":[]},{"propositions":[],"lastnames":["Barzilay"],"firstnames":["Regina"],"suffixes":[]}],"title":"Language Understanding for Text-based Games Using Deep Reinforcement Learning","journal":"arXiv.org","volume":"cs.CL","number":"","pages":"","year":"2015","abstract":"In this paper, we consider the task of learning control policies for text-based games. In these games, all interactions in the virtual world are through text and the underlying state is not observed. The resulting language barrier makes such environments challenging for automatic game players. We employ a deep reinforcement learning framework to jointly learn state representations and action policies using game rewards as feedback. This framework enables us to map text descriptions into vector representations that capture the semantics of the game states. We evaluate our approach on two game worlds, comparing against a baseline with a bag-of-words state representation. Our algorithm outperforms the baseline on quest completion by 54% on a newly created world and by 14% on a pre-existing fantasy game.","location":"","keywords":"","bibtex":"@Article{Narasimhan2015,\nauthor = {Narasimhan, Karthik and Kulkarni, Tejas and Barzilay, Regina}, \ntitle = {Language Understanding for Text-based Games Using Deep Reinforcement Learning}, \njournal = {arXiv.org}, \nvolume = {cs.CL}, \nnumber = {}, \npages = {}, \nyear = {2015}, \nabstract = {In this paper, we consider the task of learning control policies for text-based games. In these games, all interactions in the virtual world are through text and the underlying state is not observed. The resulting language barrier makes such environments challenging for automatic game players. We employ a deep reinforcement learning framework to jointly learn state representations and action policies using game rewards as feedback. This framework enables us to map text descriptions into vector representations that capture the semantics of the game states. We evaluate our approach on two game worlds, comparing against a baseline with a bag-of-words state representation. Our algorithm outperforms the baseline on quest completion by 54\\% on a newly created world and by 14\\% on a pre-existing fantasy game.}, \nlocation = {}, \nkeywords = {}}\n\n\n","author_short":["Narasimhan, K.","Kulkarni, T.","Barzilay, R."],"key":"Narasimhan2015","id":"Narasimhan2015","bibbaseid":"narasimhan-kulkarni-barzilay-languageunderstandingfortextbasedgamesusingdeepreinforcementlearning-2015","role":"author","urls":{},"downloads":0},"bibtype":"article","biburl":"https://gist.githubusercontent.com/stuhlmueller/a37ef2ef4f378ebcb73d249fe0f8377a/raw/6f96f6f779501bd9482896af3e4db4de88c35079/references.bib","creationDate":"2020-01-27T02:13:34.225Z","downloads":0,"keywords":[],"search_terms":["language","understanding","text","based","games","using","deep","reinforcement","learning","narasimhan","kulkarni","barzilay"],"title":"Language Understanding for Text-based Games Using Deep Reinforcement Learning","year":2015,"dataSources":["hEoKh4ygEAWbAZ5iy"]}