Knowledge management: the issue of multimedia contents

Knowledge Management is a very important topic in business and in academy research. There are many fields of applications for knowledge management, including cognitive science, sociology, management science, information science, \nknowledge engineering, artificial intelligence, and economics. Many studies on different aspects of Knowledge Management have been published, becoming common in the early 1990s. In this work, we want to represent Knowledge through a mixed-iterative approach, where top-down and bottom-up analyses of the knowledge domain \nwhich has to be represented are applied: these are typical approaches for this kind of problems. In this case, they are applied following an iterative approach \nwhich allows, through further refinements, for the efficient formalization able to represent the domain's knowledge of interest. We start from the concept of the “domain knowledge base”. The fundamental \nbody of knowledge available on a domain is the knowledge valuable for the knowledge users. We need to represent and manage this knowledge, to define a formalization and codification of the knowledge in the domain. After this \nformalization we can manage this knowledge using knowledge repositories. In this thesis, we present four different formalization and management of knowledge for multimedia contents, using our proposed approach: \n1. User Generated Contents from famous platform (Flickr, YouTube, etc.); \n2. audio recordings regarding linguistic corpus and information added to that \ncorpus with annotations; \n3. knowledge associated with construction processes; \n4. descriptions and reviews of Italian wines. \nThe most important result we achieved with this thesis was the opportunity to make this disparaged knowledge available and manageable. In the current market, exploiting existing knowledge is a mainstream business, but in order to \nexploit it, one must be able to manage it first. As a token of this importance, not only about ten scientific publications, but most of all a number of industrial \nresearch projects, in partnership with ICT companies – one of which with a total value above one million Euros – stemmed from the studies discussed in this thesis.