Our main focus in Machine Learning is the integration of knowledge modelling, case-based reasoning and inductive learning methods.

  • The NOOS representation language provides a computational framework for designing and implementing knowledge systems with learning capabilities.
  • Inductive learning with INDIE, a bottom-up relational learning system for Noos, an object-centered representation language.
  • Learning in Multiagent Systems:

    Our focus is on systems of agents that cooperate and learn. The agents have "competence models" that allow them to learn how to cooperate. The long term goal is to develop labour division in Multiagent Systems as an effect of the interplay of cooperation and learning.

Foundations of Analogical Inference and their Applications to Symbolic Reasoning and Learning

Initial/final date: 
14 maig 1993 a 13 maig 1996
Main researcher: 
Project type: 
-
Funding Entity: 
TIC 93-0122
Description: 
Funding Amount (€): 
0.00
Research line: 
Integration of Problem Solving and Learning
Acronym: 
ANALOG

Multi-Agent Context-Sensitive Adaptive Planning System

Initial/final date: 
01 novembre 2002 a 31 octubre 2005
Main researcher: 
Project type: 
-
Funding Entity: 
TIC2002-04146-C05-01
Description: 
Currently, society begins to see some new hardware devices, such as the Personal Digital Assistants (PDAs), mobile phones with increasing computing capabilities, or, even, new programmable freezers or washing machines with network connections. In this type of computation, the context that surrounds people has the ability of recognising us, exchanging information with us, and adapting to our needs. From a scientific point of view, this new computing paradigm poses numerous and very important questions for the eventual development of commercial applications of this technology. There are already some mature technologies that, from a standalone point of view, have shown to be successful in related domains, such as Internet, in relation to their ability to connect people with each other and with the distributed information. Examples of these technologies within the field of Artificial Intelligence are multi-agent systems, planning (and scheduling), machine learning, or user modelling. However, the future generation of viable commercial applications of interest to people requires the analysis how these techniques can be integrated within a single tool or system, so that it would be possible to develop industrial applications of ubiquitous computing. Up to now, very few researchers in the world have focused on this type of research, basically due to the need of expertise within the same group in a very diverse and disperse set of techniques. The main objective of the project is the analysis, design and implementation of a multi-agent system with the ability to perform hierarchical, temporal and resource planning and scheduling in the area of ubiquitous computing. The system will also be dynamic in that it will be able to learn from past problem solving experiences, as well as automatically acquiring a user model. To show the viability of the approach in a specific domain, we will focus on an application of e-tourism within a given city.
Funding Amount (€): 
0.00
Research line: 
Integration of Problem Solving and Learning
Acronym: 
SAMAP