Research activity in 1997
In 1997, the group has been active on several topics: a subgroup worked on the integration of numerical and symbolic learning techniques to learn and refine First Order Logic knowledge bases. This work led to the development of a refinement system (FONN) based on connectionist learning strategies. Another subgroup continued the study of Genetic Algorithms and developed a new system (CG-NET) that implements a distributed stochastic search algorithm to learn First Order Logic knowledge bases. From the theoretical viewpoint, convergency properties of FONN and properties of stochastic operators have been widely investigated.
The group headed by Francesco Bergadano, conducted extensive researches on data mining. In particular, two application domains were studied: the Adaptive System Management (ASM) and the Adaptive Network Management (ANM). In both fields, a huge set of parameters are gathered over a long period of time about the system or network under observation. At regular intervals (varying from a few seconds to a few minutes), a snapshot of the status of the system (network) is recorded. The whole information is then given in input to a learning system, in order to synthesize logical rules able to predict the evolution of the system (network) load in the near future. For the learning process the systems Regal and ReliC (currently under development) were used, achieving a classification accuracy varying from 90% for the ASM domain to 70% for the ANM.
As concern the national projects, a subgroup has been involved in a collaboration with CSELT (the research labs of Telecom Italia) aimed at developing a tool for building a symbolic characteristic description of data, mined from a large database.
1997 Publications
M. Baldoni, C. Baroglio, D. Cavagnino and G. Lo Bello: "Extraction of Discriminant Features from Image Fractal Encoding", In Advances in Artificial Intelligence LNAI 1321, pages 127-138, Springer Verlag, 1997.
C. Baroglio, A. Giordana, G. Lo Bello. and R. Piola: "Learning Fuzzy Controllers from Examples", in Statistics and Machine Learning: The Interface, C. Taylor and R. Nakhaeizadeh (eds.), Wiley & Sons, 1997.
C. Baroglio: "Exploiting qualitative knowledge to enhance skill acquisition", in Proc. of ECML-97, European Conference on Machine Learning, pages 49-56, (Prague, Czech Republic, 1997).
M. Botta, A. Giordana and R. Piola: "FONN: Combining First Order Logic with Connectionist Learning", in Proc. of the 14th Int. Conf. on Machine Learning, (Nashville, TN, 1997), pp. 48-56.
M. Botta, A. Giordana and R. Piola: "Refining Numerical Terms in Horn Clauses", In Advances in Artificial Intelligence LNAI 1321, pages 13-23, Springer Verlag 1997.
M. Botta, A. Giordana and R. Piola: "Refining First Order Theories with Neural Networks", in Proc. of the 10th ISMIS, LNAI 1325, pages 84-93, 1997.
A. Giordana, F. Neri, L. Saitta and M. Botta: "Integrating Multiple Learning Strategies in First Order Logics", Machine Learning Journal, vol. 27, n. 3, 1997, pp .209-240.
F. Neri, L. Saitta and A. Tiberghien: "Modelling physical knowledge acquisition in children with machine learning", in Proc. of 19th Annual Conference of the Cognitive Science Society. (London, UK, 1997), Lawrence Erlbaum Associates, pp. 566–571.
F. Neri, L. Saitta and A. Tiberghien: "Modelling Conceptual Change: An Interdisciplinary Approach", In Advances in Artificial Intelligence LNAI 1321, pages 1-12, Springer Verlag 1997.
L. Saitta, F. Neri and A. Tiberghien: "World model construction in children during physics learning", in Proc. of the 10th ISMIS, LNAI 1325, (Charlotte, NC, 1997), pp. 127-136.
Lorenza Saitta has been co-director, in the period 1994-1997, of the European Project "Learning in Humans and Machines", funded by the ESF. Also Filippo Neri participated to the project.
Lorenza Saitta is one of editor of the Machine Learning Journal published by Kluwer.
During the period 1993-1997, the Machine Learning Group has also been one of the 10 main nodes of the CEE European Network of Excellence in Machine Learning (MLNet), in which Lorenza Saitta has been Convener of the Research Technical Committee.
Francesco Bergadano is in the Editorial Board of the Machine Learning Journal published by Kluwer.
