KNOWLEDGE-BASED SYSTEMS

PI: José Julio Alferes

The NOVA LINCS Knowledge-Based Systems (KBS) group conducts research in the broad area of knowledge representation and reasoning. Building upon the long-standing tradition of the first Portuguese group for artificial intelligence, it currently hosts 11 FTE members (including 2 postdocs) and 14 PhD students.

With the dramatic increase in the amount of data available on the global web, it is no longer enough to be able to search it and organize it so that humans can examine and understand its content. In order to extract more value from data and make an intelligent use of it, we need to move from data to meaning, and ultimately allow machines to reason with that data using processes similar to human deductive reasoning and inference. This would also enable new classes of knowledge-rich software applications and services, supporting dynamic data and application integration, automation and interoperation of business processes, and problem-solving in various domains.

The Linked Data and the Semantic Web initiatives have led the way in developing key technologies to realize a Web of Knowledge. In the KBS group we investigate logic-based languages for representing/modeling data equipped with semantically rich information about its meaning, and develop efficient reasoning methods and tools that can make sense of joint heterogeneous knowledge. Our results are being exploited in cooperation with the industry, one example being the development of an application to integrate large amounts of data (from satellites, sensors, web, etc.) and produce risk indicators for known agricultural pests.

Our work on representation languages and reasoning methods is not confined to applications in the larger scale of the Web. We have employed our techniques in Multi-Agent Systems, to investigate the dynamics of social cooperation, in Social Networks, to model argumentation and debates, and in assisting scientific and engineering disciplines, namely to allow declarative specifications of model-based design, model-based experiments and simulation, which are increasingly replacing or enhancing physical experiments and prototypes, making use of distributed heterogeneous data-sets. In this context we develop efficient constraint base modeling and solving tools, which we exploit in concrete applications such as inverse engineering and reliability, structural bioinformatics, and oceanography.

The KBS Group is the national leading research group in its scientific area, with an internationally recognized high profile. Among its members we find two ECCAI Fellows and one of the Founders of Logic Programming – a distinction awarded by the Association of Logic Programming to 15 pioneers in the field.

We are members of the editorial board of important journals such as ACM Comp. Surveys, TPLP, J. of Applied Logic, and regularly belong to the program committee of the top conferences in the area (e.g. IJCAI, KR, ECAI, AAAI, ISWC, AAMAS, ICLP, CP), where we also regularly publish.