IFIP TC12 - Artificial Intelligence

Decision Making from Big Data Sets & Streams 

There is little doubt that the digital infrastructure of the next decade will look radically different from today’s. All sorts of products and devices will be connected to the Internet and to each other via ultra-wide band 4G (and, soon, 5G) mobile networks. Massive streams of data from digital devices will go global, along with other types of social media and rich business data. According to many, the possibility of "going full data" (i.e. handling the entire event/data streams generated by people and organizations' behavior, as opposed to sampling them to obtain traditional datasets) has the potential to dramatically improve the quality of decision making, creating a wealth of business opportunities. However, the "full data" option is not straightforward. The talk will cover some important aspects of it. First, it will provide a clear understanding of when full data is better than sampling. Secondly, it will present techniques to perform the "semantic lifting" needed  to bring events' (and context) representation at the level of abstraction suitable for specific decision making goals. Thirdly, the talk will review available architectural choices and component toolkits for data processing, as well as data integration, interoperability and trustworthiness standards.