Improving Generalization Performance in Co-Evolutionary Learning

被引:34
作者
Chong, Siang Yew [1 ,2 ]
Tino, Peter
Ku, Day Chyi [4 ]
Yao, Xin [3 ]
机构
[1] Univ Nottingham, Sch Comp Sci, Semenyih 43500, Malaysia
[2] Univ Nottingham, Sch Comp Sci, Automated Scheduling Optimizat & Planning Res Grp, Nottingham NG8 1BB, England
[3] Univ Birmingham, Sch Comp Sci, Ctr Excellence Res Computat Intelligence & Applic, Birmingham B15 2TT, W Midlands, England
[4] Multimedia Univ, Fac Informat Technol, Cyberjaya 63100, Malaysia
基金
英国生物技术与生命科学研究理事会; 英国工程与自然科学研究理事会;
关键词
Co-evolutionary learning; evolutionary computation; generalization; iterated prisoner's dilemma; machine learning; Othello; PRISONERS-DILEMMA; NEURAL-NETWORKS; COEVOLUTION; DIVERSITY; GAME;
D O I
10.1109/TEVC.2010.2051673
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Recently, the generalization framework in co-evolutionary learning has been theoretically formulated and demonstrated in the context of game-playing. Generalization performance of a strategy (solution) is estimated using a collection of random test strategies (test cases) by taking the average game outcomes, with confidence bounds provided by Chebyshev's theorem. Chebyshev's bounds have the advantage that they hold for any distribution of game outcomes. However, such a distribution-free framework leads to unnecessarily loose confidence bounds. In this paper, we have taken advantage of the near-Gaussian nature of average game outcomes and provided tighter bounds based on parametric testing. This enables us to use small samples of test strategies to guide and improve the co-evolutionary search. We demonstrate our approach in a series of empirical studies involving the iterated prisoner's dilemma (IPD) and the more complex Othello game in a competitive co-evolutionary learning setting. The new approach is shown to improve on the classical co-evolutionary learning in that we obtain increasingly higher generalization performance using relatively small samples of test strategies. This is achieved without large performance fluctuations typical of the classical approach. The new approach also leads to faster co-evolutionary search where we can strictly control the condition (sample sizes) under which the speedup is achieved (not at the cost of weakening precision in the estimates).
引用
收藏
页码:70 / 85
页数:16
相关论文
共 45 条
[1]  
[Anonymous], P 2006 IEEE S COMP I
[2]  
[Anonymous], 2001, EXERCISES PROBABILIT
[3]  
[Anonymous], 1987, Genetic Algorithms and Simulated Annealing
[4]   The effect of tag recognition on non-local adaptation [J].
Ashlock, D ;
Powers, B .
CEC2004: PROCEEDINGS OF THE 2004 CONGRESS ON EVOLUTIONARY COMPUTATION, VOLS 1 AND 2, 2004, :2045-2051
[5]   Fingerprinting: Visualization and Automatic Analysis of Prisoner's Dilemma Strategies [J].
Ashlock, Daniel ;
Kim, Eun-Youn .
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION, 2008, 12 (05) :647-659
[6]   Understanding representational sensitivity in the iterated prisoner's dilemma with fingerprints [J].
Ashlock, Daniel ;
Kim, Eun-Youn ;
Leahy, Nicole .
IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART C-APPLICATIONS AND REVIEWS, 2006, 36 (04) :464-475
[7]  
Axelrod R., 1984, EVOLUTION COOPERATIO
[8]  
Back T., 1997, IEEE Transactions on Evolutionary Computation, V1, P3, DOI 10.1109/4235.585888
[9]   Evolution, neural networks, games, and intelligence [J].
Chellapilla, K ;
Fogel, DB .
PROCEEDINGS OF THE IEEE, 1999, 87 (09) :1471-1496
[10]   Measuring generalization performance in coevolutionary learning [J].
Chong, Siang Yew ;
Tino, Peter ;
Yao, Xin .
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION, 2008, 12 (04) :479-505