INVITED SPEAKERS

 

 

 

 

Marco Gori (Dipartimento di Ingegneria dell'Informazione - Università di Siena)

 

 

 

Diffusion learning by prior and acquired links

 

A remarkable limitation of most supervised approaches to machine learning is that they need a lot of labeled examples to attain a satisfactory generalization to new examples. In order to face this relevant problem, a number of methods have been devised that operate under the framework of transductive or semi-supervised learning. Starting from these premises, I this talk I show the crucial role of links, either given in advance or properly acquired by unsupervised learning, for learning in huge collections of data (e.g. in the Web). I introduce the notion of diffusion learning, whose computational scheme is very much related to Web ranking algorithms based on link analysis. Diffusion learning is based on supervision that involves both the content and the links. The diffusion mechanisms are proposed in a unified view, in which the prior information expressed by the links turns out to spread evidence through the entire data collection, regardless of the presence of supervision. I give very promising experimental results on the learning of functions used in link analysis, like PageRank, and show the effectiveness of the proposed theory especially when the rank depends jointly on the page content and on the links. Finally, I argue that the propagation of the relationships expressed by the links reduces dramatically the sample complexity with respect to traditional learning machines operating on vector spaces, thus opening the doors to the solution of remarkable large scale real-world problems involving ranking, classification, and link prediction.

 

 

 

Marco Gori received the Ph.D. degree in 1990 from Università di Bologna, Italy. From October 1988 to June 1989 he was a visiting student at the School of Computer Science (McGill University, Montreal). In 1992, he became an Associate Professor of Computer Science at Università di Firenze and, in November 1995, he joint the Università di Siena, where he is currently full professor of computer science. His main interests are in machine learning, with applications to pattern recognition, Web mining, and game playing. He is especially interested in the formulation of relational machine learning schemes in the continuum setting. He is the leader of the WebCrow project for automatic solving of crosswords that has recently outperformed human competitors in an official competition taken place within the ECAI-06 conference. He is co-author of the book "Web Dragons: Inside the myths of search engines technologies," Morgan Kauffman (Elsevier), 2006. Dr. Gori serves (has served) as an Associate Editor of a number of technical journals related to his areas of expertise, including IEEE Transaction on Neural Networks, Pattern Recognition, Neural Networks, Neurocomputing, Pattern Analysis and Application, the International Journal of Document Analysis and Recognition, and the International Journal on Pattern Recognition and Artificial Intelligence. He has been the recipient of best paper awards and keynote speakers in a number of international conferences. He was the Chairman of the Italian Chapter of the IEEE Computational Intelligence Society and the President of the Italian Society for Artificial Intelligence. He is a fellow of the ECCAI and of the IEEE.

 

 

 

 

 

Massimo Paolucci (DoCoMo Euro-labs)

 

 

 

On bringing AI to the street level

 

We are living in a world in which digital services are increasingly provided at the “street level”. Such services which include information services such as public transport information, payment systems, and ticketing are becoming one of the media through which we interact with the environment in which we live. In this trend, mobile phones are the tool through which we gain access to these services, and through them to operate in the real world. Artificial intelligence has developed a great deal of theories on how computers can deal autonomously with the real world out there. Such theories ranging from planning to learning and knowledge representation can play a crucial role in the development of mobile computing. In this talk I will discuss our experience with developing a mobile platform to deploy on mobile phones that for intelligent service provisioning and I will try to highlight the emerging challenges for the different areas of AI.

 

 

 

Massimo Paolucci is a senior researcher at DOCOMO Euro-labs where he is conducting research on service provisioning and composition to mobile users exploiting web services, and semantic web technology. Before joining DOCOMO he worked on semantic web services and agent technology at Carnegie Mellon University. Massimo has been a member of the UDDI technical committee and a member of the OWL-S coalition. He served in the organizing committee of ISWC as PC member is a number of AI related conferences including AIMSA, IJCAI, AAMAS, ISWC and numerous workshops.