Take a look at the new version of this page: [ beta version ]. Tell us what you think.

ABSTRACT

Neighbor search is a fundamental task in machine learning, especially in classification and retrieval. Efficient nearest neighbor search methods have been widely studied, with their emphasis on data structures but most of them did not consider the underlying global geometry of a data set. Recent graph-based semi-supervised learning methods capture the global geometry, but suffer from scalability and parameter tuning problems. In this paper we present a (nearest) neighbor search method where the underlying global geometry is incorporated and the parameter tuning is not required. To this end, we introduce deterministic walks as a deterministic counterpart of Markov random walks, leading us to use the minimax distance as a global dissimilarity measure. Then we develop a message passing algorithm for efficient minimax distance calculation, which scales linearly in both time and space. Empirical study reveals the useful behavior of the method in image retrieval and semi-supervised learning.

REFERENCES

• 1: Mikhail Belkin , Partha Niyogi, Laplacian Eigenmaps for dimensionality reduction and data representation, Neural Computation, v.15 n.6, p.1373-1396, June 2003 [doi>10.1162/089976603321780317]
• 2: Mikhail Belkin , Partha Niyogi, Semi-Supervised Learning on Riemannian Manifolds, Machine Learning, v.56 n.1-3, p.209-239 [doi>10.1023/B:MACH.0000033120.25363.1e]
• 3: Carroll, J. D. (1995). Minimax length links of a dissimilarity matrix and minimum spanning trees. Psychometrika, 60, 371--374.
• 4: Hong Chang , Dit-Yan Yeung, Locally linear metric adaptation for semi-supervised clustering, Proceedings of the twenty-first international conference on Machine learning, p.20, July 04-08, 2004, Banff, Alberta, Canada [doi>10.1145/1015330.1015391]
• 5: Delalleau, O., Bengio, Y., & Roux, N. L. (2005). Efficient non-parametric function induction in semi-supervised learning. Proceedings of International Workshop on Artificial Intelligence and Statistics (pp. 96--103).
• 6: Carlotta Domeniconi , Jing Peng , Dimitrios Gunopulos, Locally Adaptive Metric Nearest-Neighbor Classification, IEEE Transactions on Pattern Analysis and Machine Intelligence, v.24 n.9, p.1281-1285, September 2002 [doi>10.1109/TPAMI.2002.1033219]
• 7: Francois Fouss , Alain Pirotte , Marco Saerens, A Novel Way of Computing Similarities between Nodes of a Graph, with Application to Collaborative Recommendation, Proceedings of the 2005 IEEE/WIC/ACM International Conference on Web Intelligence, p.550-556, September 19-22, 2005 [doi>10.1109/WI.2005.9]
• 8: Garcke, J., & Griebel, M. (2005). Semi-supervised learning with sparse grids. Proceedings of ICML, Workshop on Learning with Partially Classified Training Data (pp. 19--28).
• 9: Gower, J. C., & Ross, G. J. S. (1969). Minimum spanning tree and single-linkage cluster analysis. Applied Statistics, 18, 54--64.
• 10: Kschischang, F. R., Frey, B. J., & Loeliger, H. A. (2001). Factor graphs and the sum-product algorithm. IEEE Trans. Information Theory, 47, 498--519.
• 11: Roweis, S. T., & Saul, L. (2000). Nonlinear dimensionality reduction by locally linear embedding. Science, 290, 2323--2326.
• 12: Gregory Shakhnarovich , Trevor Darrell , Piotr Indyk, Nearest-Neighbor Methods in Learning and Vision: Theory and Practice (Neural Information Processing), The MIT Press, 2006
• 13: Daniel A. Spielman , Shang-Hua Teng, Nearly-linear time algorithms for graph partitioning, graph sparsification, and solving linear systems, Proceedings of the thirty-sixth annual ACM symposium on Theory of computing, June 13-16, 2004, Chicago, IL, USA [doi>10.1145/1007352.1007372]
• 14: Szummer, M., & Jaakkola, T. (2002). Partially labeled classification with Markov random walks. Advances in Neural Information Processing Systems (pp. 945--952). MIT Press.
• 15: Tenenbaum, J. B., de Silva, V., & Langford, J. C. (2000). A global geometric framework for nonlinear dimensionality reduction. Science, 290, 2319--2323.
• 16: Fei Wang , Changshui Zhang, Label propagation through linear neighborhoods, Proceedings of the 23rd international conference on Machine learning, p.985-992, June 25-29, 2006, Pittsburgh, Pennsylvania [doi>10.1145/1143844.1143968]
• 17: Xing, E. P., Ng, A. Y., Jordan, M. I., & Russel, S. (2003). Distance metric learning, with application to clustering with side-information. Advances in Neural Information Processing Systems. MIT Press.
• 18: Zhou, D., Bousquet, O., Lal, T. N., Weston, J., & Schöölkopf, B. (2004). Learning with local and global consistency. Advances in Neural Information Processing Systems (pp. 321--328). MIT Press.
• 19: Zhou, D., & Schöölkopf, B. (2004). Learning from labeled and unlabeled data using random walks. Pattern Recognition, Proceedings of the 26th DAGM Symposium (pp. 237--244).
• 20: Zhu, X., Ghahramani, Z., & Lafferty, J. (2003). Semi-supervised learning using Gaussian fields and harmonic functions. Proceedings of International Conference on Machine Learning (pp. 912--919).
• 21: Xiaojin Zhu , John Lafferty, Harmonic mixtures: combining mixture models and graph-based methods for inductive and scalable semi-supervised learning, Proceedings of the 22nd international conference on Machine learning, p.1052-1059, August 07-11, 2005, Bonn, Germany [doi>10.1145/1102351.1102484]