Evolving a population code for multimodal concept learning

We describe an evolutionary method for learning concepts of objects from multimodal data. The proposed method uses a population code (hypernetwork representation), i.e. a collection of codewords (hyperedges) and associated weights, which is adapted by evolutionary computation based on observations of positive and negative examples. The goal of evolution is to find the best compositions and weights of hyperedges to estimate the underlying distribution of the target concepts. We discuss the relationship of this method with estimation of distribution algorithms (EDAs), classifier systems, and ensemble learning methods. We evaluate the method on a suite of image/text benchmarks. The experimental results demonstrate that the evolutionary process successfully discovers salient codewords representing multi-modal feature combinations for describing and distinguishing different concepts. We also analyze how the complexity of the population code evolves as learning proceeds.

[1]  Yoav Freund,et al.  Boosting a weak learning algorithm by majority , 1990, COLT '90.

[2]  Qingfu Zhang,et al.  On the convergence of a class of estimation of distribution algorithms , 2004, IEEE Transactions on Evolutionary Computation.

[3]  Jason H. Moore,et al.  Learning classifier systems: a complete introduction, review, and roadmap , 2009 .

[4]  Pedro Larrañaga,et al.  Estimation of Distribution Algorithms , 2002, Genetic Algorithms and Evolutionary Computation.

[5]  John H. Holland,et al.  Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence , 1992 .

[6]  J. A. Lozano,et al.  Estimation of Distribution Algorithms: A New Tool for Evolutionary Computation , 2001 .

[7]  Cyrus Rashtchian,et al.  Every Picture Tells a Story: Generating Sentences from Images , 2010, ECCV.

[8]  David G. Lowe,et al.  Object recognition from local scale-invariant features , 1999, Proceedings of the Seventh IEEE International Conference on Computer Vision.

[9]  Marin Ferecatu,et al.  Semantic interactive image retrieval combining visual and conceptual content description , 2007, Multimedia Systems.

[10]  R. Manmatha,et al.  Automatic image annotation and retrieval using cross-media relevance models , 2003, SIGIR.

[11]  Heikki Mannila,et al.  Prediction with local patterns using cross-entropy , 1999, KDD '99.

[12]  A. Doucet,et al.  A Tutorial on Particle Filtering and Smoothing: Fifteen years later , 2008 .

[13]  Y. Mori,et al.  Image-to-word transformation based on dividing and vector quantizing images with words , 1999 .

[14]  Heiko Wersing,et al.  Towards autonomous bootstrapping for life-long learning categorization tasks , 2010, The 2010 International Joint Conference on Neural Networks (IJCNN).

[15]  Marco Tomassini,et al.  Experimental Study of Multipopulation Parallel Genetic Programming , 2000, EuroGP.

[16]  Erick Cantú-Paz,et al.  A Survey of Parallel Genetic Algorithms , 2000 .

[17]  Qingfu Zhang,et al.  Structure learning and optimisation in a Markov-network based estimation of distribution algorithm , 2009, 2009 IEEE Congress on Evolutionary Computation.

[18]  P. Strevens.  Iii , 1985 .

[19]  Byoung-Tak Zhang,et al.  Hypernetworks: A Molecular Evolutionary Architecture for Cognitive Learning and Memory , 2008, IEEE Computational Intelligence Magazine.

[20]  Ralph R. Martin,et al.  A Sequential Niche Technique for Multimodal Function Optimization , 1993, Evolutionary Computation.

[21]  Kwang Ryel Ryu,et al.  A Dual-Population Genetic Algorithm for Adaptive Diversity Control , 2010, IEEE Transactions on Evolutionary Computation.