High Performance Computational Intelligence. We are focused on the development of new CI models and algorithms with good scaling properties. The algorithms are designed for all the leading HPC arquitectures are considered for this algorithms. On the one hand, for processing large datasets both clusters of computers as well as cloud computing platforms, with distributed applications based on MPI. And, on the other hand, to reach the highest performance levels the most effective processors are exploited: multi-core micro processors, GPUs and Intel Xeon Phi coprocessors.

Big Data
The design of effective procedures to handle large datasets is challenging tasks. An effective approach for this is the MapReduce approach, for which Hadoop is the most popular open-source implementation. We develop new models in Hadoop, most of which integrates seamlessly in Mahout.

High quality scientific and enginnering software.
We consider the implementation of algorithms and procedures as an end of the utmost importance themselves and not just intermediate steps towards evaluating hypothesis. That is why we intend to develop high quality implementation in terms of complete packages or libraries which will help researchers and practitioners in two goals:

• Boosting advances in research by provided robust implementations
• Helping toward reproducible research

Time series analysis and forecasting. Since they are present in most area of human knowledge, time series are an important data set to deal with. Computational Intelligence offers effective tools which can cooperate with methods from other areas to enhance the analysis of time series and to produce more accurate forecasters. We are interested in investigation in the deep connections between CI models and other areas models, as well as creating more effective forecasters.

Applications of the methods to real-world problems:
Railway maintenance optimization, biometrics, windmill diagnosis and prognosis, financial data mining and processing, expert systems for Medicine, ...