Using data and natural language processing to make the most of Europe’s cultural heritage
‘The teams behind the apps’ is a series of videos made by the EU Datathon finalists. Get to know this year’s teams and their apps over 12 episodes, twice per week, in the run-up to the competition finals. On 20 October, the teams will pitch their apps to the jury that will select the winners. The audience will also have a chance to support their favourite team in the Public Choice Award vote. Register now
The UNIOR NLP team, Maria Pia di Buono, Gennaro Nolano, Giulia Speranza and Johanna Monti, joined us for a short interview in which we talked about the app, the idea that led to its creation and the decision to participate in the EU Datathon.
How are you doing with the app? What’s the status?
We are late, of course! However, last-minute panic is activating our creativity. We are pushing on with the development of all the modules needed for our real-time chatbot, and we plan to have a working prototype by the beginning of October (i.e. 2 hours before the EU Datathon final event).
At this stage, we are refining the natural-language understanding and information extraction modules, which allow Maggie to understand users’ requests and retrieve the data that will be used to provide answers. The last step will be to develop the natural-language generator module, which also includes data aggregation and merging, and a visualisation procedure which will offer users a friendly and intuitive way to engage with cultural content.
As an academic research group, we usually only create proofs of concept to test our research hypotheses, which means that working on a complete app has been quite challenging for us, but we will see where it takes us!
How did you come up with the idea for this app?
The app is the result of years of research in this area, as our group has always been committed to making cultural content easily accessible for all users by developing systems and applications for the cultural heritage domain. In fact, our research team has already exploited European open data (mainly data from the Europeana portal) in several experiments aimed at improving the current state-of-the-art in natural-language processing tasks.
We strongly believe that developing systems that are capable of managing cultural content is a way to support pluralism, cultural growth, inclusiveness and equality. In this respect, dialogue systems and conversational agents can contribute to easing users’ access to knowledge according to their information needs, reawakening their curiosity and interest in cultural heritage.
Besides being the virtual assistant we would definitely use ourselves, Maggie offers us a new way to appreciate and enjoy the incredible European cultural legacy.
Why did you decide to take part in the EU Datathon?
‘Because it’s there’ (as a famous mountain climber said) and because it’s exciting! The EU Datathon is the mountain we are climbing to exit the comfort zone of academia made of papers and proofs of concept. We are racing with ourselves, and we feel that getting involved in a new, stimulating and challenging environment, such as the EU Datathon, is the best way to get things done.
When we found out about the competition, we felt that it would be a great opportunity to bring together all our previous efforts and develop an app to give people easy access to European cultural content in today’s digital age.
Finally, we think that the Datathon is giving us the possibility to show how natural-language processing methodologies can be a valuable means to improve the accessibility and use of open data.
To find out more about UNIOR NLP, watch this 1-minute video.