This use case was led by Italian broadcasters RAI, and it targeted the delivery of personalised news services. Guided mainly by speech-to-text, news programs are segmented into news items and enriched by text medadata relating to names, places, topics, or generic concepts. These are matched to, and filtered by, user preferences and delivered via alerts (in the case of breaking news), an automatically generated local news channel, or by user-initiated search.
Showcase goals
- To exploit the flexible/adaptive end-to-end architecture developed in NoTube
- To verify the actual feasibility of a system for the creation and the delivery of a set of local personalised news services based on the repurposing and enrichment of news item material already transmitted over generic broadcast streams
- RAI also used the built demonstrators and pilots to analyse, both from the service provider and the final user point of view, new personalised services and new models of interactive programmes, initially news-oriented, outside the regular programming
These objectives were part of the overall NoTube project goals.
Research areas covered in this showcase
- Common metadata exchange format
- Interfacing with legacy systems
- Multilingual semantic enrichment of content based on entity recognition
- Recommendations based on Linked Data
- Social TV APIs
- Personalisation of content following user behaviour
- Automatic device adaptation
Who was involved? The lead partner for this Use Case was RAI working in collaboration with Polymedia and other project partners in the consortium.
Personalized Semantic News service
In this prototype, the front-end application is designed to run on a large touch screen. A touchscreen PC emulates a close viewing experience such as personally watching on a tablet. For the more traditional “far away” viewing experience at home the interface can equally be reproduced using pointing devices (like Wii) or gesture recognition (like Kinect).
The user can access the service selecting its personal icon or the authentication could be integrated with the ambient intelligence inside the Home Ambient.
When the user enters inside the service he can start to see the personalized news program that has been automatically edited following his own preferences; these can be dynamically extracted following the behaviour of the user in the social network or can be fixed by the user himself through the selection of its favourite genres. In this prototype only the fixed generation of preferences is considered. Watching his personalized newscast, the user can choose to look at it in a traditional broadcast oriented way or to use the PVR oriented enhanced functionalities, such as the pause or the skip functionality for each single news.
The next figure shows the news items title list up on the right of the screen which can be used to navigate through the personalized news program.
The prototype allows the viewing of an animated tag cloud which displays concepts that are considered significant inside the news, all concepts are mapped with DBpedia and clicking on one of them the relevant DBpedia abstract will be displayed at the bottom of the screen. The default language for the abstract is Italian but if it doesn’t exist then the English one is displayed. In particular, each concept on the screen “pulse” in a different way so that higher relevance tags are more “clickable/touchable” than the others.
In the prototype the NIC structure allows the presence of external related links, for each news item a set of URLs is available for the navigation in web sites of other media companies that expose information about the same news. The user can navigate through the related external links and touching an item from the list the application opens an integrated browsing window showing the related news.
When the browsing window is opened the video behind is automatically set to pause, on the contrary when the window is closed the video plays again automatically. Inside the application the Facebook activity stream of the user is integrated. The user can interact with friends as he could do in the original Facebook home page.
Another Facebook functionality which has been integrated is the chat, visible on the left of the screen in the next figure.
The service allows to share personalized news items with all Facebook friends and it’s also possible to receive suggestions from people who are inside the NoTube environment connected to the social network. In particular next figure shows how a NoTube user can send a News Item suggestion with a joined personal message to friends and how this suggestion is notified to the Home Facebook GUI of a friend, while the last figure shows how the same friend could receive the suggestion on his socially connected NoTube environment.
To give more background, take a look at our poster:
Personalized Semantic News on mobile devices
In the updated prototype, the application has been implemented onto a mobile device, with both integration with the main TV screen and coordinated play-out of the video in the mobile screen now possible.
The WP7a 3rd prototype allows the final user to access the Personalized Newscast Program using the Smartphone as a secondary screen. A full demo of the new prototype can be seen in the below slides and video:
RAI Personalised Semantic News demo, third prototype 2011
The mobile video cropping technology is provided by Thomson Video Networks.
In particular, the reframed video of News items is used in the 3rd prototype of WP7a for the play out of a particular news inside the Smartphone. When the final user reaches a particular interested news he can choose if play it on the TV screen or to watch the news directly on the Smartphone screen.
Find out more
Related publications and presentations can be found in the Things to read section.