TD-gammon revisited: integrating invalid actions and dice factor in continuous action and observation space

Download

index.pdf

Date

2018

Author

Usta, Engin Deniz

Metadata

Show full item record

Item Usage Stats

239
views

94
downloads

After TD-Gammon's success in 1991, the interest in game-playing agents has risen significantly. With the developments in Deep Learning and emulations for older games have been created, human-level control for Atari games has been achieved and Deep Reinforcement Learning has proven itself to be a success. However, the ancestor of DRL, TD-Gammon, and its game Backgammon got out of sight, because of the fact that Backgammon's actions are much more complex than other games (most of the Atari games has 2 or 4 different actions), the huge action space has much invalid actions, and there is a dice factor which involves stochasticity. Last but not least, the professional level in Backgammon has been achieved a long time ago. In this thesis, the latest methods in DRL will be tested against its ancestor game, Backgammon, while trying to teach how to select valid moves and considering the dice factor.

Subject Keywords

Backgammon., Computer games., Machine learning., Artificial intelligence.

URI

http://etd.lib.metu.edu.tr/upload/12622619/index.pdf
https://hdl.handle.net/11511/27576

Collections

Graduate School of Natural and Applied Sciences, Thesis

Suggestions

OpenMETU
Core

Transformative Impacts of High-Speed Railway (HSR) Stations on Urban Space: The Case of Ankara HSR Station Songulen, Nazli; Memluk, Nihan Oya (2014-05-11) Urban transformation became one of the premises on the urban agenda in Turkey, which was accelerated with the introduction of the neo-liberal policies in the 1980s and became apparent in the 2000s as rapid and massive urban transformations mainly held by macro-scaled public investments. In this regard, high-speed railway (HSR) systems, newly introduced to the Turkish case, appear as one of these macro-scale public investments. In this context, the European literature on HSR development projects reveals that...
Clean energy transition in the Turkish power sector: A techno-economic analysis with a high-resolution power expansion model Kat, Bora (2023-06-01) The Turkish power sector achieved rapid growth after the 1990s in line with economic growth and beyond. However, domestic resources did not support this development and therefore resulted in a high dependency on imported fossil fuels. Furthermore, the governments were slow off the mark in introducing policies for increasing the share of renewable energy. Even late actions of the governments, as well as significant decreases in the cost of wind and especially solar technologies, have recently brought the Tur...
Cross-cultural Perspectives of Successful Aging: Young Turks and Europeans Cosco, Theodore D.; Brehme, David; Grigoruta, Nora; Kaufmann, Lisa-Katrin; Lemsalu, Liis; Meex, Ruth; Schuurmans, Angela A. T.; Sener, Neslihan; Stephan, Blossom C. M.; Brayne, Carol (2015-11-02) Successful aging (SA) has been conceptualized in a number of ways. Despite increasing research into how laypersons define SA, few studies capturing lay perspectives of SA in younger cohorts and in non-English speaking countries have been undertaken. The current study examines cross-cultural perspectives of SA in young (aged 18-35), lay adults from a variety of continental European countries and Turkey. Participants were recruited via snowball sampling from social network sites and invited to participate in ...
Analyzing RD activities of foreign enterprises in emerging economies. Lessons from Turkey Erdil, Erkan; Pamukcu, Mehmet Teoman (Science And Technology Policies Research Center, Middle East Technical University (Ankara, Turkey), 2011-01-01) Emerging economies have played an important role in the internationalization of R&D activities at least since the 1990s. Turkey, an emerging economy and at same time an accession country to the European Union which signed a Customs Union Agreement with the EU already in 1995, is no exception. In-depth face-to-face semi-structured interviews were conducted with R&D directors of 26 multinational companies operating in Turkey –with headquarters located in France, Germany, Italy, Japan, Switzerlandand USA- in t...
Labor market transformation in technologically renewed firms: printing and publishing sector in case of Istanbul Erdoganaras, Fatma Çetinkaya; Ersoy, Melih; Department of City and Regional Planning (2002) In both in the World as a whole and Turkey, the printing and publishing sector is one of the sectors which has experienced a restructuring process after 1980. This was largely realized depending on new technology as a result of the recent developments in transportation, communication and telecommunication technologies and particularly intensive use of the computer aided production system. The introduction of new technologies brings about several impacts on a sector depending on social, economic, political a...

Citation Formats

E. D. Usta, “TD-gammon revisited: integrating invalid actions and dice factor in continuous action and observation space,” M.S. - Master of Science, Middle East Technical University, 2018.