A Theory on Big Data

Abstract

CLAIRE the ‘initiative of a pan-EU confederation of AI Research Labs’[Cl18] anticipates a humane AI which is based on ethical and trustworthy tenets empowering citizens and society. Hence in order to achieve a human AI, new semantic categories of standards must be written enabling stakeholders to implement a responsible AI. Communication however generates lots of unstructured data sets to be classified and structured into data types. AI-based algorithms are suitable to derive - from data sets and data types - more sophisticated and implicitly given information that can further be enriched to knowledge about autonomous communicating processes or even autonomous behaving humans. The knowledge gained from applying data enrichment algorithms in turn may be used for reasoning and prediction purposes, thus improving applications a lot. The aim of this project is to find a common unified tool able to handle big data flows and types that allows the recognition of hidden information from both data flows and data types. The hidden information is sometimes also called meta-data, thus being implicitly existing but by tools becoming explicitly seen.