Browse our site
CENTRIA researchers focus on two main sub-areas: Knowledge Represenation and Reaoning (KRR) and Soft Computing and Constraints (SCC).
Knowledge Representation and Reasoning Representing knowledge in forms suitable for automatic processing and developing inference algorithms to manipulate the knowledge have long been cornerstones within Artificial Intelligence, and together are a core concern to intelligent applications.
Research in KRR at CENTRIA ranges from foundational aspects to applications, including the development and implementation of knowledge-based systems.
Our foundational research focuses primarily on Computational Logics approaches to the representation of knowledge and to reasoning methods. In addition, our research covers representation and reasoning with imprecise probabilities, foundations of mathematical logic, and several topics in philosophical logic.
Our applications of knowledge-based systems focus on the Semantic Web and Multi-Agent Systems (mostly adopting Logic Programming as an implementation vehicle), but also include automated proof systems and formal models in cognitive science.
In summary, CENTRIA’s specific contributions to KRR range from foundational work on non-monotonic reasoning and evolution of knowledge bases, knowledge updates and belief revision, reasoning about actions and causality, answer-set programming, the combination of ontology and rule based knowledge, Bayesian epistemology, imprecise probabilities, implementations for the web and semantic web, and of multi-agent systems.
Soft Computing and Constraints (SCC) Soft Computing (SC) models reasoning processes which are tolerant of imprecision, uncertainty, and partial truth. Our SC research adopts a variety of techniques to this effect, including machine learning, neural networks, fuzzy systems, evolutionary computation, and Bayesian networks. Constraint Programming (CP) models reasoning which exploits hard constraints to problem solving through propagation, global constraints, symmetry breaking, and heuristics.
Complex problems often require hybrid approaches that rely on heuristics and soft constraints to approximate SC techniques. CENTRIA's SCC research exploits the complementarity of SC and CP approaches and applies them to areas such as biology, medicine, social sciences, management sciences, and similar fields.
The main contributions of CENTRIA to the SC field are fuzzy techniques and machine learning, with specific applications to oceanography; evolutionary models to understand the emergence of phenomena such as cooperation or ethic norms in complex agent societies; foundations of probability theory and causal Bayesian networks.
In CP our main contributions are novel domains and constraint propagation techniques, and their application to Bioinformatics problems; extending constraints on continuous domains to dynamic phenomena and probabilistic reasoning; and generic constraint solvers, with efficient software architectures and exploiting hardware parallelism.