The Intelligent Systems Research Line aims to study, develop and apply computational systems capable of solving problems that require intelligence to be solved. It aims to apply the concepts, techniques and tools of Artificial Intelligence to help solve conceptual and practical problems in computing and other areas of knowledge, in addition to studying and developing state-of-the-art artificial intelligence that aims its practical use in industrial and agro-industrial processes. This line includes systems that evolve and adapt (Evolutionary Systems), distributed intelligence systems (Multi-Agent Systems) and systems that are capable of learning from experience (Machine Learning).
The area of Evolutionary Systems proposes an alternative paradigm to traditional mathematical/physical/computational models. This new paradigm does not require prior knowledge of a method to find the solution to the problem, as it is based on evolutionary mechanisms such as self-organization and adaptive behavior, which are recurrent. It is also understood as a branch of natural computing that includes the topics of artificial life, fractal geometry, complex systems and computational intelligence.
In the Machine Learning area, we study algorithms capable of improving its performance through experience, allowing us to build systems that are capable of learning and detecting concepts, categories and patterns. Such techniques are essential to model solutions that do not have a known algorithmic form and whose data set is too large to be analyzed by people.
In the Knowledge Representation and Reasoning area, mechanisms are investigated to represent, organize and search for the interoperability of information about characteristics of an environment, enabling their to be useed by computer systems to support the resolution of complex problems. Among its application areas, we highlight the use of ontology engineering and data science applied to education, medicine, bioinformatics and robotics.
In the area of Recommendation Systems, methods for recommending objects (people, documents, practices, among others) are researched considering the profile, interests and context of users based on their characteristics and information from different sources. It seeks to understand the interests, objectives and characteristics of these users, aiming at the development of personalized solutions that assist in the decision-making processes. Among its application areas we highlight smart educational environments, health and wellness, marketing and advertising.
Finally, in the Multiagent Systems area, we seek to solve problems that are inherently distributed through the social action of autonomous agents. Thus, the solution of the problems occurs through the performance of all these agents following mechanisms of coordination, cooperation, conflict resolution, communication, etc. A large number of problems in the industry are distributed and therefore must be treated in this way to enable efficient solutions with real applicability.
- Cristian Cechinel
- Marilton Sanchotene de Aguiar
- Paulo Roberto Ferreira Jr.
- Ricardo Matsumura Araújo
- Tiago Thompsen Primo