The university is a hub of new ideas with experts of different domains regularly sharing their knowledge and research findings through seminars, public lectures, presentations, etc. Each year, hundreds of talks are held in a university. These seminars can be seen as invaluable sources to gain multi-disciplinary cross-subject knowledge. However, seminar announcements are normally only circulated within closed mailing-lists, groups of research teams and departments and it is often difficult for students or even academics to pick up information of all the talks that they may be interested in. This project proposes to design and implement a prototype of a web/mobile-based system that integrates and updates seminar/talk announcement in the university and recommends seminar announcements to registered users using a range of AI/Machine learning algorithms.
A web/mobile-based system that are able to extract talk abstracts from multiple information sources, e.g., seminar mailing-lists of different departments, collect them into a database and perform a range of analytics such as semantic vectorisation, classification, filtering, as an automated recommender system for registered users. The input to the system will be texts of title/abstract/speaker of talks to be held in the university, as well as user attributes such as degree, subjects, research interests, etc.
Lab allocations have not been finalised