Effectiveness of considering state similarity for reinforcement learning

Date

2006-01-01

Author

Girgin, Sertan
Polat, Faruk
Alhajj, Reda

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

63
views

0
downloads

This paper presents a novel approach that locates states with similar sub-policies, and incorporates them into the reinforcement learning framework for better learning performance. This is achieved by identifying common action sequences of states, which are derived from possible optimal policies and reflected into a tree structure. Based on the number of such sequences, we define a similarity function between two states, which helps to reflect updates on the action-value function of a state to all similar states. This way, experience acquired during learning can be applied to a broader context. The effectiveness of the method is demonstrated empirically.

URI

https://hdl.handle.net/11511/54252

Journal

INTELLIGENT DATA ENGINEERING AND AUTOMATED LEARNING - IDEAL 2006, PROCEEDINGS

Collections

Department of Computer Engineering, Article

Suggestions

OpenMETU
Core

Positive impact of state similarity on reinforcement learning performance Girgin, Sertan; Polat, Faruk; Alhaj, Reda (Institute of Electrical and Electronics Engineers (IEEE), 2007-10-01) In this paper, we propose a novel approach to identify states with similar subpolicies and show how they can be integrated into the reinforcement learning framework to improve learning performance. The method utilizes a specialized tree structure to identify common action sequences of states, which are derived from possible optimal policies, and defines a similarity function between two states based on the number of such sequences. Using this similarity function, updates on the action-value function of a st...
Effectiveness of 5E learning cycle model on hihg school students' understanding of solubility equilibrium concept Aydemir, Nurdane; Geban, Ömer; Department of Secondary Science and Mathematics Education (2012) The purpose of this study was to investigate the effect of instruction based on 5E learning cycle model (LCI) compared to Traditional Instruction (TI) and gender on 11th grade students’ understanding of solubility equilibrium concept, students’ perceived motivation, use of learning strategies, and attitudes towards chemistry. There were 53 students in the experimental group instructed by the LCI and 56 students in the control group instructed by the TI. Solution Concept Test and Science Process Skills Test ...
Effectiveness of context based approach through 5E learning cycle model on students' understanding of chemical reactions and energy concepts and their motivation to learn chemistry Çiğdemoğlu, Ceyhan; Geban, Ömer; Department of Secondary Science and Mathematics Education (2012) The aim of study was to investigate the effect of context-based approach (CBA) through 5E learning cycle (LC) model over traditional instruction on students’ understanding, achievement, and chemical literacy on chemical reactions and energy concepts. The effect of instruction on students’ motivation to learn chemistry and the factors of motivation questionnaire were also explored. Additionally, the effect of gender difference was investigated. Six eleventh grade classes with 187 students taught by three tea...
Effect of structuring cooperative learning based on conceptual change approach on students' understanding of the concepts of mixtures and their motivation Belge Can, Hatice; Boz, Yezdan; Department of Secondary Science and Mathematics Education (2013) The purpose of this study is to investigate the effect of structuring cooperative learning based on conceptual change approach on grade nine students’ understanding the concepts of mixtures and their motivation, compared to traditional instruction. Mixtures Concept Test (MCT), self-efficacy for learning and performance, task value, control of learning beliefs, and test anxiety sub-scales of Motivated Strategies for Learning Questionnaire (MSLQ), and mastery approach goals, mastery avoidance goals, performan...
Learning by Automatic Option Discovery from Conditionally Terminating Sequences Girgin, Sertan; Polat, Faruk; Alhajj, Reda (2006-08-28) This paper proposes a novel approach to discover options in the form of conditionally terminating sequences, and shows how they can be integrated into reinforcement learning framework to improve the learning performance. The method utilizes stored histories of possible optimal policies and constructs a specialized tree structure online in order to identify action sequences which are used frequently together with states that are visited during the execution of such sequences. The tree is then used to implici...

Citation Formats

S. Girgin, F. Polat, and R. Alhajj, “Effectiveness of considering state similarity for reinforcement learning,” INTELLIGENT DATA ENGINEERING AND AUTOMATED LEARNING - IDEAL 2006, PROCEEDINGS, pp. 163–171, 2006, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/54252.