AnnoMarket – Getting the word out
While the project team have been beavering away preparing for the public beta release of the AnnoMarket platform, we’ve also been out and about spreading the word to potential customers and suppliers. Here’s a round-up of this autumn’s activities.
In September Valentin Tablan and I spoke about AnnoMarket at the Meta-Forum in Berlin. This is an annual conference on language technologies, and this year’s event had a special emphasis on big data text analytics and multilingual web services in Europe. AnnoMarket was introduced as part of a session dedicated to the needs of European industry. The marketplace’s cloud-based infrastructure and growing range of non-English language resources (such as entity extractors) could help businesses needing to do multilingual text annotation in a quick and cost-effective way.
Early October saw the KESW (Knowledge Engineering and Semantic Web) conference in St Petersburg, Russia. For the first time in the event’s history the proceedings were entirely in English rather than Russian, in a bid to open up the event to more practitioners. The bulk of the attendees though were from Russia or Eastern Europe, so it was a good opportunity to promote AnnoMarket’s offerings in languages from these countries. I did a five minute “lightning talk” (it’s not easy introducing AnnoMarket in such a short time!) and also took part in the poster session, where I met a number of active GATE or text analytics users. I got some valuable feedback about how users would like to interact with the platform, for example, being able to use sliders or filters to select the best app for their needs.
The next appointment was back in London, at Shoreditch’s famous Town Hall for the “Newshack” event organised by BBC News Labs. I was part of the AnnoMarket team with Ian Roberts, Mark Greenwood and Dominic Rout (all representing the University of Sheffield). The BBC provided access to a corpus of 3000 news stories, and some APIs. We explored various ideas around using linked data to power related links, or integrating social media and news articles. Our final prototype was meant as a showcase for GATE technology and could potentially be used by journalists needing richer semantic querying capability than a keyword search could provide. We used GATE to annotate the BBC corpus and pull in related metadata from sources such as Google Maps and DBPedia. The application was queryable from a GATE Mímir front end. The query we demonstrated in the (hair-raising) two minute pitch at the end allowed us to find articles about all “fat cat” bankers who had received bonuses above a certain large amount of money (leveraging GATE’s ability to annotate quantities of money).
Finally, on 3rd December I spoke at the Women in Data meetup, at the prestigious Open Data Institute in London. Along with another PA colleague we presented an NLP night for an audience of geeks (not just female ones!) Again, my presentation was intended to elicit feedback about the marketplace, and use cases from potential stakeholders. There was a lively discussion about the platform’s APIs, how potential suppliers could interact with the beta and also whether a future iteration of AnnoMarket could offer different models for integrating with private clouds.
We’ve got lots more planned for 2014 so keep an eye on this site and on our Twitter feed!