Data Analysis and Applications

The Data and Applications Analysis line encompasses both aspects related to data structure, storage and location of data, as well as the process of producing information and extracting knowledge from data. The research in this line involves data structures and algorithms to support the process of extracting knowledge from data, as well as in the application and development of functions for pre-processing, data analysis and data post-processing. Along this line, it is appropriate to establish a datacentric treatment for these processes, which commonly correspond to in-silico experiments.

This line of research has great synergy with the domain of the problem addressed and has a highly multidisciplinary character. The application areas associated with the line are commonly observed in the governmental, scientific and business areas. In the governmental axis, applications in the areas of planning, health, transport, education and communication can be highlighted. In the scientific area, applications in astronomy, seismic, health and pharmaceuticals stand out. In the business area, the development of software for the Information Technology market and its applications in marketing, social networks, multimedia systems and smart cities stand out. All of these domains bring with them different technologies present in the area of ​​production and consumption of large volumes of data that can be exemplified from the Internet of Things (IoT), sensor networks and interactive applications.

In this deluge of data, the particular characteristics and demands of applications bring numerous challenges in managing large volumes of data. With regard to the processing of large volumes of data, there is an urgent need to use high performance processing (PAD) to enable large-scale data analysis. There are important challenges in establishing these analyzes, commonly modeled as workflows. In these workflows, activities and data are directed to execution in some PAD environment (e.g., clusters, grids, clouds).

The different applications, while presenting themselves as applied research, often provide the opportunity to develop new theoretical frameworks in basic research, of a more general character, for the solution of these practical problems. Basic research occurs both when establishing representations that are agnostic to the environment in which they will be performed, as well as in the development of new functions that are more appropriate to the treatment and extraction of knowledge from different data sources and that have a wide spectrum of use.

Machine Learning and Optimization

The activities of this line of research involve the study, the proposal, the development of computational models in optimization and scientific computing, as well as the application of machine learning techniques in the description of complex problems from Computer Science to several domains, such as Biology, Engineering, Finance, Health and Logistics.

In general, the search for solutions can be divided into two paths, methods that use information obtained through theoretical studies and methods that use historical data to “learn” how to deal with future scenarios, the so-called data-driven models.

In the first path, we have Computational Modeling that encompasses Optimization and Scientific Computing, which are areas that seek to develop tools to support decision making or assist in the understanding of real phenomena. Such areas use mathematical models and the development of efficient algorithms to solve theoretical and applied problems, as well as the use of high-performance computing techniques. Among the techniques of interest in this area, we can highlight Computational Intelligence techniques such as Genetic Algorithms, Ant Colonies, among others. In the context of applications, the problems involving Smart Cities, Industry 4.0 and Internet of Things stand out, but are not restricted.

Belonging to the second path, Machine Learning (AM), subarea of ​​Artificial Intelligence, investigates how machines can learn some task after receiving a set of experiences (in the form of historical data). Two sub-areas of AM are of particular interest to our research group: Deep Learning and Reinforced Learning. In these areas we are interested in how to simulate the behavior of the human brain, as well as in the development of techniques that allow the rational behavior of software agents. Of interest in this area are problems involving pattern recognition, computer vision, natural language processing, as well as task planning and scheduling, games and many others.

The multidisciplinary nature of this line of research allows researchers and students involved to acquire solid and broad training due to its applicability in several domains and knowledge of the various facets of the problems addressed.