Option and constraint generation using Work Domain Analysis

Tokadli, Guliz; Feigh, Karen M.

doi:10.1109/smc.2014.6974232

Cited by 4 publications

(10 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The agent should avoid taking those actions defined as constraints. RL methods are used to generate policies for how an algorithm should make a decision or play a game in terms of an agent interacting with an environment (Tokadlı & Feigh, 2014;Mohan & Laird, 2010).…”

Section: Reinforcement Learningmentioning

confidence: 99%

“…Recent work by Tokadlı and Feigh (2014) has investigated using the Abstraction Hierarchy modeling from Work Domain Analysis as an additional complementary method to elicit domain knowledge for use in machine learning algorithms. In that study, the authors used the results of AHs and translated the players' domain knowledge into options and constraints.…”

Section: Introductionmentioning

confidence: 99%

“…In that study, the authors used the results of AHs and translated the players' domain knowledge into options and constraints. Options and constraints were defined with the hierarchy that is appropriate for reinforcement learning algorithm schema (Tokadlı & Feigh, 2014).…”

Section: Introductionmentioning

confidence: 99%

“…Specifically, human domain experts were interviewed, immediately following a familiarization period playing the game; then, AHs were created for each individual and composite hierarchies were made for the best and worst performing players in each game. The hierarchies were used to identify the player's knowledge of the work domain and strategies which the players described and applied, for a ML agent which are composed of options and constraints as documented in (Tokadlı & Feigh, 2014).…”

Section: Introductionmentioning

confidence: 99%

“…It also describes the generation of AHs for Mario Bros. and how to operationalize performance-based AHs for RL algorithms. The generation of AHs for Pac-Man and generation of options and constraints sets are provided in (Tokadlı & Feigh, 2014), and is summarized under Pac-Man Study Section.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Application of Abstraction Hierarchies to Incorporate Human Knowledge for Machine Learning

Tokadli

Feigh

2015

Proceedings of the Human Factors and Ergonomics Society Annual

Self Cite

View full text Add to dashboard Cite

We evaluate the use of the Abstraction Hierarchy (AH) modeling technique from Work Domain Analysis (WDA) to generate policies suitable for use by a reinforcement learning agent. The paper extends previous work utilizing the classic arcade game, Pac-Man to the more complex world of Mario Bros. This paper presents the investigation of AHs for Mario Bros. to generate options and constraints for a reinforcement learning agent. For both games, AHs were generated for high and low performers using qualitative data. The differences between high and low performers’ AHs were translated into sets of options and constraints as suitable information into machine learning algorithms. A general AH form, which was built to enable researchers to generate an AH for simple game domains using the results of the previous study and this study, supports that an AH for another simple game might be generated using only quantitative data collected from game play and the observations of a researcher.

show abstract

Section: Reinforcement Learningmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations