On Wednesday 6th February the team (in the form of Simon) were down in Sheppard’s Bush to attended the Semantic Media @ BBC workshop event organised by Queen Mary University London and the BBC, in a venue right next to the iconic BBC Television Centre building. The event was intended to bring together interested parties from the world of academia and business to held the BBC gather and make the most of its production metadata in a semantic way.
Although the event was advertised under the ‘semantic’ banner, the majority of the speakers (all of whom were from the BBC) were largely outlining problems in artificial intelligence, and the majority of the attendees — at least those from academia — seemed to be from an AI / machine learning background.
The speakers broadly fell into two categories: the first being engineers or various types from the BBC’s London R+D facility. The second being archivists working on collecting and publishing historic information about the BBC’s culture, engineering feats and programme schedules.
Despite the lack of awareness of real (that is W3C) Semantic Web approaches, some of the problems proffered by the various BBC speakers may actually benefit in part from a linked data approach. For example, their media datasets seem to be non-trivial in size (eg. BBC Sport based in Salford allegedly gather many thousands of hours of video per year, with its associated metadata). If the correct data could be gathered and given the semantic triplestore treatment, it might form the basis of a powerful navigation and analysis tool. And when the work Ensemble has done with the Semantic Web and video is included, the prospects could be very interesting indeed.
The team has submitted a proposal for a mini-project, working to find ways to exploit some of the technologies we’ve developed in the past, and Simon has already contacted some of those at the event to suggest possible avenues for exploration.