Today, companies and individuals are immersed in interconnected digital ecosystems in which enormous amounts of heterogeneous data, the so-called Big Data, are produced at an unprecedented speed. In this scenario, the Big Data approaches and technologies are designed to make computable problems that involve large volumes of highly heterogeneous data by format and structure, often generated in real-time, which cannot be dealt with existing computer technologies.
The purpose of the lab is to study methods, models, algorithms, approaches and languages for managing and analyzing heterogeneous Big Data such as texts, images, video, networks, structured data, geo-referenced, multi-media and multi-dimensional, which vary over time, in order to promote the creation of innovative Smart Technologies. These technologies are capable of transforming Big Data into Smart Data, that is data enriched by explicit semantics, which can be modelled and extracted by semantic technologies, and by implicit semantics, obtained through machine learning techniques, artificial intelligence, and data analytics. Moreover, the definition of Smart Models with Operation Research methodologies, and mathematical evaluations for the study of the quantitative and qualitative characteristics of the data are highly relevant.
Smart technologies make more efficacious and efficient the activities of business analytics, decision support, content management and improve marketing activities, and the quality of products / services offered to their customers / users. In general, they optimize all the application aspects present in different business and operational processes of organizations in different application areas.
The lab research topics involve two macro areas: Mathematical and Smart Models and Smart Data. The first includes models and algorithms related to the transportation planning and logistics networks: non-differentiated and non-convex optimization methods; methods and models for resolving supervised and / or semi-supervised classification problems and for image processing.
In the second, models, approaches and languages for semantic data management, both in structured and unstructured formats: approaches for extracting information from the web and from documents, even in natural language; languages based on artificial intelligence and machine learning for manipulating and querying structured and unstructured data; models and techniques for managing and extracting content from social media, analysis and mining of complex networks.
The laboratory’s scientific and technological results are applied in various fields of application including healthcare, telecommunications, electronic commerce, the Internet of things, and security.