Algorithm 652: HOMPACK: a suite of codes for globally convergent homotopy algorithms
ACM Transactions on Mathematical Software (TOMS)
Text Classification from Labeled and Unlabeled Documents using EM
Machine Learning - Special issue on information retrieval
Theory of Globally Convergent Probability-One Homotopies for Nonlinear Programming
SIAM Journal on Optimization
Exploitation of Unlabeled Sequences in Hidden Markov Models
IEEE Transactions on Pattern Analysis and Machine Intelligence
IEEE Transactions on Pattern Analysis and Machine Intelligence
Learning Hidden Variable Networks: The Information Bottleneck Approach
The Journal of Machine Learning Research
Iterative Local-Global Energy Minimization for Automatic Extraction of Objects of Interest
IEEE Transactions on Pattern Analysis and Machine Intelligence
Homotopy-based semi-supervised Hidden Markov Models for sequence labeling
COLING '08 Proceedings of the 22nd International Conference on Computational Linguistics - Volume 1
Maximum entropy modeling in sparse semantic tagging
HLT-SRWS '04 Proceedings of the Student Research Workshop at HLT-NAACL 2004
Probability-one homotopy maps for tracking constrained clustering solutions
Proceedings of the High Performance Computing Symposium
Hi-index | 0.00 |
A number of modern learning tasks involve estimation from heterogeneous information sources. This includes classification with labeled and unlabeled data as well as other problems with analogous structure such as competitive (game theoretic) problems. The associated estimation problems can be typically reduced to solving a set of fixed point equations (consistency conditions). We introduce a general method for combining a preferred information source with another in this setting by evolving continuous paths of fixed points at intermediate allocations. We explicitly identify critical points along the unique paths to either increase the stability of estimation or to ensure a significant departure from the initial source. The homotopy continuation approach is guaranteed to terminate at the second source, and involves no combinatorial effort. We illustrate the power of these ideas both in classification tasks with labeled and unlabeled data, as well as in the context of a competitive (min-max) formulation of DNA sequence motif discovery.