DISI Seminar seminar
Tue Feb 14, 2012 at 11:00
Meeting room Yoram Ofek Via Sommarive, 5
Paolo Frasconi
kLog is a logical and relational language for kernel-based learning. It allows users to specify logical and relational learning problems at a high level in a declarative way. It builds on simple but powerful concepts: learning from interpretations, entity/relationship data modeling, logic programming and deductive databases (Prolog and Datalog), and graph kernels. kLog is a statistical relational learning system but unlike other statistical relational learning models, it does not represent a probability distribution directly. It is rather a kernel-based approach to learning that employs features derived from a grounded entity/relationship diagram. These features are derived using a novel technique called graphicalization}: first, relational representations are transformed into graph based representations; subsequently, graph kernels are employed for defining feature spaces. kLog can use numerical and symbolic data, background knowledge in the form of Prolog or Datalog programs (as in inductive logic programming systems) and several statistical procedures can be used to fit the model parameters. The kLog framework can -- in principle -- be applied to tackle the same range of tasks that has made statistical relational learning so popular, including classification, regression, multitask learning, and collective classification.
About the speaker Paolo Frasconi is a professor of Computer Science at the University of Florence. His research interests are in the area of machine learning, with particular emphasis on algorithms for structured and relational data, and applications to bioinformatics. He is an Associate Editor of the Artificial Intelligence Journal and an Action Editor of the Machine Learning Journal. He co-chaired PAIS 2012, the AAAI 2010 Special Track on AI and Bioinformatics, the 20th Int. Conf. on Inductive Logic Programming 2010, and the 5th International Workshop on Mining and Learning with Graphs (2007).
Contact: Andrea Passerini