A History Tree Heuristic to Generate Better Initiation Sets for Options in Reinforcement Learning

Date

2016-09-02

Author

DEMİR, ALPER
Cilden, Erkin
Polat, Faruk

Metadata

Show full item record

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

Item Usage Stats

29
views

0
downloads

Options framework is a prominent way to improve learning speed by means of temporally extended actions, called options. Although various attempts focusing on how to derive high quality termination conditions for options exist, the impact of initiation set generation of an option is relatively unexplored. In this work, we propose an effective heuristic method to derive useful initiation set elements via an analysis of the recent history of events.

URI

https://hdl.handle.net/11511/35681

DOI

https://doi.org/10.3233/978-1-61499-672-9-1644

Collections

Department of Computer Engineering, Conference / Seminar

Suggestions

OpenMETU
Core

GENERATING EFFECTIVE INITIATION SETS FOR SUBGOAL-DRIVEN OPTIONS DEMİR, ALPER; Cilden, Erkin; Polat, Faruk (World Scientific Pub Co Pte Lt, 2019-03-01) Options framework is one of the prominent models serving as a basis to improve learning speed by means of temporal abstractions. An option is mainly composed of three elements: initiation set, option's local policy and termination condition. Although various attempts exist that focus on how to derive high-quality termination conditions for a given problem, the impact of initiation set generation is relatively unexplored. In this work, we propose an effective goal-oriented heuristic method to derive useful i...
A process capability based assessment model for software workforce in emergent software organizations TANRIÖVER, ÖMER ÖZGÜR; Demirörs, Onur (Elsevier BV, 2015-01-01) Software process improvement frameworks for software organizations enable to identify opportunities for improving the processes as well as establishing road maps for improvement. However, software process improvement practice showed that to achieve a sustained, leveraged state, software organizations need to focus on the workforce as much as the process. Software process improvement frameworks address the people dimension indirectly through processes. To complement process assessment models/methods, there i...
A Rule-based domain specific language for fault management Kaya, Özgür; Doğru, Ali Hikmet; Department of Computer Engineering (2014) A fault management framework has been developed where a rule-based event processing language is also developed that provides improvement to the existing approaches in terms of time responsiveness. Reference architectures were developed for the fault management domain including fault avoidance capabilities. Such capability is for taking precautionary actions before the fault happens, while most of the fault tolerance techniques are intended for detecting a fault after it happens, hence utilizing the time wit...
A Reflexion Model based Architecture Conformance Analysis Toolkit for OSGi-compliant Applications Cilden, Evren; Oğuztüzün, Mehmet Halit S. (2017-04-07) Component-based software platforms like OSGi facilitate the development of complex software. As software systems become more complicated, tool support is often a necessity for assuring the conformance between designed and implemented architectures. We present ARTOS, an architecture toolkit to facilitate the design and conformance analysis of the software running on the OSGi platform. The toolkit consists of an architecture editor and a conformance analyzer. The editor provides definition constructs specific...
A multitone model of complex enveloped signals and its application in feedforward circuit analysis Coskun, AH; Mutlu, Ayşe Ceyda; Demir, S (Institute of Electrical and Electronics Engineers (IEEE), 2005-06-01) Analytical tools that characterize nonlinear systems are essential and need to be developed for initial rapid optimizations and understanding of the system performance. Modeling of the input signal is a crucial part of this task. In this paper, a multitone representation for an arbitrary double-side banded (symmetric spectrum around carrier frequency) stochastically not well-defined signal and its application to a feedforward circuit, which involve two nonlinear amplifiers, couplers, phase, and delay units ...

Citation Formats

A. DEMİR, E. Cilden, and F. Polat, “A History Tree Heuristic to Generate Better Initiation Sets for Options in Reinforcement Learning,” 2016, vol. 285, Accessed: 00, 2020. [Online]. Available: https://hdl.handle.net/11511/35681.