Statistical Relational Learning (SRL) is an emerging sub-field of machine learning that deals with machine learning in relational domains where observation variables may be missing, partially observed, and/or noisy. Traditionally, relational and logical reasoning, probabilistic and statistical reasoning, and machine learning are research fields in their own rights. Nowadays, they are becoming increasingly intertwined. Applications within e.g. bioinformatics, transportation systems, social network analysis, citation analysis, and robotics provide uncertain information about varying numbers of entities and relationships among the entities. Traditional machine learning approaches are able to cope either with uncertainty or with relational representations but typically not with both.

The research group STREAM focuses on all aspects of SRL. Supported by the »Fraunhofer Attract« program, the group develops techniques for statistical learning and acting within worlds of objects and relations among these objects. Tasks and techniques investigated include alignment of rich objects, visualizations and embeddings of rich data, lifted inference, (relational) Gaussian processes, and (relational) reinforcement learning, among others. The ultimate goal is to implement these techniques for real-world applications.

Visitors

• Bernd Gutmann, K. U. Leuvem, Belgium, Oct. 27-30, 2009.
• Ingo Thon, K. U. Leuven, Belgium: Jul. 28-29, 2009
• Scott Sanner, NICTA-SML, Australia: Jun. 24-29, 2009
• Novi Quadrianto, NICTA-SML, Australia: Feb. 21-23, 2009
• Keiichi Horio, Kyushu Institute of Technology, Japan: Dec. 16, 2008 - Feb. 14, 2009
• Christian Plagemann, Stanford University, USA: Nov. 26-27, 2008
• Sriraam Natarajan, University of Wisconsins at Madison, USA: Sept. 5-9, 2008; Jun. 23 - Jul. 1, 2009  

 Press

• Koelnische Rundschau, Oct. 31, 2009
• Koelner Stadt-Anzeiger, Oct. 30, 2009
• Koelner Express, Oct. 29, 2009.

Project Members

• Dipl.-Inform. Mirwaes Wahabzada
• Dr. Kristian Kersting
• Dipl.-Inform. Fabian Hadiji
• Dipl.-Inform. Babak Ahmadi
• Dr. Zhao Xu

Publications

• K. Kersting, Z. Xu. Learning Preferences with Hidden Common Cause Relations. In W. Buntine, M. Grobelnik, D. Mladenic, J. Shawe-Taylor, editor(s), Proceedings of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML PKDD 09), LNAI, Bled, Slovenia, Sept. 7-11 2009. Springer. [ draft | wordle | BibTeX ]
• H. Schulz, K. Kersting, A. Karwath. ILP, the Blind, and the Elephant: Euclidean Embedding of Co-Proven Queries. In L. De Raedt, editor(s), Proceedings of the 19th International Conference on Inductive Logic Programming (ILP-09), LNCS, Leuven, Belgium, July 2-4 2009. Springer. (short paper, to appear). [ draft | BibTeX ]
• M. Neumann, K. Kersting, Z. Xu, D. Schulz. Stacked Gaussian Process Learning. In H. Kargupta, W. Wang, editor(s), Proceedings of the 9th IEEE International Conference on Data Mining (ICDM-09), Miami, FL, USA, Dec. 6-9 2009. [ draft | wordle | BibTeX ]
• C. Thurau, K. Kersting, C. Bauckhage. Convex Non-Negative Matrix Factorization in the Wild. In H. Kargupta, W. Wang, editor(s), Proceedings of the 9th IEEE International Conference on Data Mining (ICDM-09), Miami, FL, USA, Dec. 6-9 2009. [ draft | wordle | BibTeX ]
• N. Quadrianto, K. Kersting, M. Reid, T. Caetano, W. Buntine. Kernel Conditional Quantile Estimation via Reduction Revisited. In H. Kargupta, W. Wang, editor(s), Proceedings of the 9th IEEE International Conference on Data Mining (ICDM-09), Miami, FL, USA, Dec 6-9 2009. (Short Paper). [ draft | BibTeX ]
• B. Ahmadi, M. Hadjieleftheriou, T. Seidl, D. Srivastava, S. Venkatasubramanian. Type-Based Categorization of Relational Attributes. Proc. of the International Conference on Extending Database Technology (EDBT), Saint-Petersburg, Russia, March 23-26 2009. [ draft | BibTeX ]
• S. Joshi, K. Kersting, R. Khardon. Generalized First Order Decision Diagrams for First Order Markov Decision Processes. In C. Boutilier, editor(s), Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI-09), Pasadena, CA, USA, July 11-17 2009. [ draft | BibTeX ]
• K. Kersting, B. Ahmadi, S. Natarajan. Counting Belief Propagation. In A. Ng, J. Bilmes, editor(s), Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence (UAI-09), Montreal, Canada, June 18-21 2009. [ draft | code | wordle | BibTeX ]
• Z. Xu, V. Tresp, A. Rettinger, K. Kersting. Social Network Mining with Nonparametric Relational Models. In H. Zhang, M. Smith, L. Giles, J. Yen, editor(s), Advances in Social Network Mining and Analysis, LNCS, Springer, 2009. [ draft | BibTeX ]
• Z. Xu, K. Kersting, V. Tresp. Multi-Relational Learning with Gaussian Processes. In C. Boutilier, editor(s), Proceedings of the International Joint Conference on Artificial Intelligence (IJCAI-09), Pasadena, CA, USA, July 11-17 2009. [ draft | wordle | BibTeX ]
• K. Kersting, K. Driessens. Non-Parametric Policy Gradients: A Unified Treatment of Propositional and Relational Domains. In A. McCallum, S. Roweis, editor(s), Proceedings of the 25th International Conference on Machine Learning (ICML 2008), Helsinki, Finland, July 5-9 2008. [ draft | BibTeX ]
• K. Kersting. SRL without Tears: An ILP Perspective. In F. Zelezny, N. Lavrac, editor(s), Proceedings of the 18th International Conference on Inductive Logic Programming (ILP-08), Number 5194 in LNCS, page 2. Prague, Czech Republic, September 10-12 2008. Springer. [ BibTeX ]
• O. Missura, K. Kersting, T. Gärtner. Towards Engaging MDPs. In A. Botea, C. Linares López, editor(s), Working Notes of the ECAI-08 Workshop on Artificial Intelligence in Games (AIG-08), Patras, Greece, July 22 2008. [ BibTeX ]
• B. Gutmann, A. Kimmig, K. Kersting, L. De Raedt. Parameter Learning in Probabilistic Databases: A Least Squares Approach. In W. Daelemans, B. Goethals, K. Morik, editor(s), In Proceedings of the European Conference on Machine Learning and Knowledge Discovery (ECML/PKDD-08), volume 5211 of LNCS, page 473-488. Antwerp, Belgium, September 15-19 2008. Springer. [ draft | BibTeX ]
• Z. Xu, K. Kersting, V. Tresp. Gaussian Process Models for Colored Graphs. In E. Airoldi, D. Blei, J. Hofman, T. Jebara, E. Xing, editor(s), Working Notes of the NIPS-08 Workshop on Analyzing Graphs, 2008. [ draft | BibTeX ]
• C. Plagemann, K. Kersting, W. Burgard. Non-stationary Gaussian Process Regression using Point Estimates of Local Smoothness. In W. Daelemans, B. Goethals, K. Morik, editor(s), In Proceedings of the European Conference on Machine Learning and Knowledge Discovery (ECML/PKDD-08), volume 5212 of LNCS, page 203-219. Antwerp, Belgium, September 15-19 2008. Springer. [ draft | BibTeX ]
• S. Natarajan, H. Bui, P. Tadepalli, K. Kersting, W.-K. Wong. Logical Hierarchical Hidden Markov Models for Modeling User Activities. In F. Zelezny, N. Lavrac, editor(s), Proceedings of the 18th International Conference on Inductive Logic Programming (ILP-08), page 192-209. Prague, Czech, September 10-12 2008. Springer. [ draft | BibTeX ]
• A. Karwath, K. Kersting, N. Landwehr. Boosting Relational Sequence Alignment. In F. Giannotti, D. Gunopulos, editor(s), Proceedings of IEEE International Conference on Data Mining (ICDM-08), Pisa, Italy, Dec. 15-19 2008. (Short Paper). [ draft | BibTeX ]