The PGMs4SDA research project

Probabilistic graphical models (PGMs) have experienced a remarkable development during the last years and have shown up as valuable tools in disciplines as Artificial Intelligence and Statistics. In the last years, much attention has been put on the use of PGMs in data mining tasks, specially in situations endowed with uncertainty. From the current state-of-the-art the natural step forward is to provide them with the ability of operating in big data contexts.

The purpose of this project is to generate a set of new methodological developments in the field of PGMs sufficiently sound and innovative as to place them as reference tools in the area of big data, through scalable data analytics. Also,the project is intended to produce the necessary software tools as to allow the development of applications based on a web-services architecture for PGMs. In this way, we expect to create a framework where mobile hardware devices could be used in big data contexts, as the core processing would rely on a centralized server that would process the data and run the necessary algorithms, while the mobile device would interact through the web services interface. Hence, the purpose of this project is twofold, by first generating new knowledge of the highest scientific quality within the field of scalable data analytics, and later on by paving the way to a fruitful technology transfer making use of the planned software platform.

Partners: