Journalism is in crisis, but new information and communication technologies (ICT) offer new opportunities. Journalists today have access to a wealth of digital information from news aggregators, social media, and open data providers in addition to traditional sources. These sources can be automatically analysed, organised, prepared, and stored with increasing semantic precision and connectedness. Theories and techniques from artificial intelligence and machine learning can be used to classify, label, cluster, detect events, and otherwise process news streams in meaningful ways. Our research group is collaborating with a software developer of news production tools for the international market. Together we have developed News Hunter, a prototype tool pipeline that analyses news and social media messages and represents them semantically in a knowledge graph that is used to classify, label, and identify clusters of messages, potentially in real time. As in similar tools, the focus is on surface similarity: pieces of information are related if they are about similar things, or if they are treated similarly by similar people. While useful for many purposes, the focus on surface similarity may drown other important and interesting connections hidden in the information. News Angler will therefore extend and build on News Hunter to support deep and innovative information mining that goes under the surface. We want to leverage the semantic news graph to identify interesting and unexpected connections between unfolding news events and prepared background information represented in the graph. To achieve this, News Angler will explore adapting, combining, and extending theories from analogical and other types of computational reasoning. The theories and techniques that News Angler will develop potentially have importance beyond journalism, as an alternative to the surface similarity-based search and recommendation services that shape the information bubbles we live in today.
Project leader: Andreas Lothe Opdahl
Institution: Institutt for informasjons- og medievitenskap