Missouri S&T Scholar's Mine Research RepositoryMissouri S&T Research
print 
Title: Negative reinforcement and backtrack-points for recurrent neural networks for cost-based abduction
Author (s): Abdelbar, A.M.
El-Hemaly, M.A.
Andrews, E.A.M.
Wunsch, Donald C.
Department/Lab Affiliations: Applied Computational Intelligence Laboratory
Electrical and Computer Engineering
Keywords: backtrack-points
cost-based abduction
heuristic perturbations
high order recurrent neural networks
learning (artificial intelligence)
negative reinforcement
perturbation techniques
recurrent neural nets
Issue Date: 2005
Publisher: Institute of Electrical and Electronics Engineers
Citation: Abdelbar, A.M.; El-Hemaly, M.A.; Andrews, E.A.M.; Wunsch, D.C., II, "Negative reinforcement and backtrack-points for recurrent neural networks for cost-based abduction" IJCNN '05. Proceedings. 2005 IEEE International Joint Conference on Neural Networks, pp. 827- 832 vol. 2, 31 July-4 Aug. 2005
Abstract: Abduction is the process of proceeding from data describing a set of observations or events, to a set of hypotheses which best explains or accounts for the data. Cost-based abduction (CKA) is an AI formalism in which evidence to be explained is treated as a goal to be proven, proofs have costs based on how much needs to be assumed to complete the proof, and the set of assumptions needed to complete the least-cost proof are taken as the best explanation for the given evidence. In this paper, we introduce two techniques for improving the performance of high order recurrent networks (HORN) applied to cost-based abduction. In the backtrack-points technique, we use heuristics to recognize early that the network trajectory is moving in the wrong direction; we then restore the network state to a previously-stored point, and apply heuristic perturbations to nudge the network trajectory in a different direction. In the negative reinforcement technique, we add hyperedges to the network to reduce the attractiveness of local-minima. We apply these techniques on a 300-hypothesis, 900-rule particularly-difficult instance of CBA.
Type: Article - Conference proceedings
text
Copyright Notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.
FULL COPYRIGHT INFORMATION:
http://www.ieee.org/web/publications/rights/policies.html
Publisher URL:
http://ieeexplore.ieee.org/iel5/10421/33090/01555959.pdf?arnumber=155595
Link to this page:
http://scholarsmine.mst.edu/post_prints/01555959_09007dcc8030d82c.html
Full Text:
01555959_09007dcc8030d831.pdf



titleNegative reinforcement and backtrack-points for recurrent neural networks for cost-based abduction
contributor.authorAbdelbar, A.M.
contributor.authorEl-Hemaly, M.A.
contributor.authorAndrews, E.A.M.
contributor.authorWunsch, Donald C.
contributor.deptlabApplied Computational Intelligence Laboratory
contributor.deptlabElectrical and Computer Engineering
subjectbacktrack-points
subjectcost-based abduction
subjectheuristic perturbations
subjecthigh order recurrent neural networks
subjectlearning (artificial intelligence)
subjectnegative reinforcement
subjectperturbation techniques
subjectrecurrent neural nets
date.issued2005
date.submitted2007
publisherInstitute of Electrical and Electronics Engineers
identifier.citationAbdelbar, A.M.; El-Hemaly, M.A.; Andrews, E.A.M.; Wunsch, D.C., II, "Negative reinforcement and backtrack-points for recurrent neural networks for cost-based abduction" IJCNN '05. Proceedings. 2005 IEEE International Joint Conference on Neural Networks, pp. 827- 832 vol. 2, 31 July-4 Aug. 2005
identifier.pub.URI
http://ieeexplore.ieee.org/iel5/10421/33090/01555959.pdf?arnumber=155595
description.abstractAbduction is the process of proceeding from data describing a set of observations or events, to a set of hypotheses which best explains or accounts for the data. Cost-based abduction (CKA) is an AI formalism in which evidence to be explained is treated as a goal to be proven, proofs have costs based on how much needs to be assumed to complete the proof, and the set of assumptions needed to complete the least-cost proof are taken as the best explanation for the given evidence. In this paper, we introduce two techniques for improving the performance of high order recurrent networks (HORN) applied to cost-based abduction. In the backtrack-points technique, we use heuristics to recognize early that the network trajectory is moving in the wrong direction; we then restore the network state to a previously-stored point, and apply heuristic perturbations to nudge the network trajectory in a different direction. In the negative reinforcement technique, we add hyperedges to the network to reduce the attractiveness of local-minima. We apply these techniques on a 300-hypothesis, 900-rule particularly-difficult instance of CBA.
typeArticle - Conference proceedings
type.DCMITypetext
type.statusFinal version
rightsThis material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.
rights.URI
http://www.ieee.org/web/publications/rights/policies.html
date.accessioned2007-04-05T14:25:17Z
date.available2007-04-05T14:25:16Z
identifier.persist.URI
http://scholarsmine.mst.edu/post_prints/01555959_09007dcc8030d82c.html
Full Text
01555959_09007dcc8030d831.pdf