Dynamic Content Monitoring and Exploration using Vector Spaces

Objectives

The aim of this project is to investigate a methodology to support specialists in monitoring, exploring and interacting with (thematic) issues in unstructured and dynamic corpora; examples are informative resource streams or corpora that involve a long time span such as set of newspaper articles, blog posts, journal paper collections, or historical document collections. The methodology will rely on representation(s) based on abstract vector spaces. The representations will be designed in order to be suitable for enrichment, generalisation, specification performed manually by the specialist or automatically by the system (e.g. exploiting profiles) though operations on vector spaces, thus providing specialists with representation tailored to their field and task. Indeed, even when considering the same issue, tailored representation can be beneficial to different specialists (e.g. historians, journalists or sociologists) for instance in terms of the level of technicality or verbosity. Moreover, issue monitoring based on different representations or issue comparison can provide useful insights for specialist analyses.

Expected Results

(i) A vector space-based methodology to represent (thematic) issues in dynamic corpora. (ii) Evidence that the quantum formalism can be beneficial to perform operations on these thematic issue representations, e.g. beneficial in terms of issue comparison or issue monitoring. (iii) An experimental system that implements the overall methodology. (iv) The design of a real case study in cooperation of specialists to experimentally evaluate the methodology.

Person in Charge

Professor Massimo Melucci, University of Padua, Italy, massimo.melucci (at) unipd.it

Expression of interest and how to apply

Click here to express your interest in this position and access details on the application procedure.