Watch with Friends launched for Glastonbury content on iPlayer, an Andy C synthetic voice robot & Misa's last week :(
Internet & Society Team
Libby’s been doing various bits of work on ±«Óătv Together, along with many others, . David’s been wireframing and designing logos for a website to showcase better images of AI, and Henry’s been trying to find artists and photographers to make these images. Tristan’s been re-writing some blogposts on understanding AI and did a talk about this for the Data Science Research Partnership. Alicia and Mathieu have been helping Galen with the next game prototype for explaining AI, built around . Chris has been organising department-wide software engineering meetings and joining meetings on and , amongst others. Henry’s been gathering people to help with our futuring process.
Data Team
It’s been a productive couple of weeks for the Data Team. In the speech-to-text stream, Ollie and Matt have been doing some painstaking detective work to understand why our new language model is causing the system to slow down, Misa has been putting the finishing touches to the team’s impressive looking paper on X-vector research and we’ve been patiently working through the various hoops the team needs to jump through to present at an upcoming conference.
In the Recs work stream, Ben has been working with Kristine to make a new dataset from a year’s worth of iPlayer viewing data. He’s also been filling in a variety of ethical and information security checklists and presenting the team's iRex work. There’s lots of interest in audio tagging at the moment, so watch this space for new outcomes from that. Polina has been working with colleagues on shaping the human values for recommendations work and doing some text processing.
In the Natural Language Processing stream, Chris has been reviewing his QMUL student’s work on automated segmentation, which is looking very promising. Chris is now giving him data from non-news segmentation data to see how that performs. He’s also been writing an app to allow him to write segmentation data, which he’s been using to process a bunch of 'From Our Own Correspondents’. Joanna has been continuing her work on finishing off our Sentiment research write-up and Andrew has been beavering away keeping the many DSRP stakeholders happy.
Finally, we’re sad to say that next week is Misa’s last week with IRFS. The whole team would like to wish her well in her new role and thank her for everything she’s bought to the department. Thanks Misa, it’s been an absolute joy :)
Interaction & Prototyping Team
The team has been working hard on completing the development of the multi-screen prototype and firming up the single screen version of the prototype with design improvements and tweaks based on internal testing. They have started the recruitment of external participants to evaluate the prototype in August and have been working with colleagues in the R&D North Lab to validate our Likert scale questions and surveys. The team has also spent some time reflecting on the direction of the team as well as scope, output and impact of the project in its current remit. Last week was also a time to celebrate Andy Conroy - who’s leaving R&D. Among all the amazing things people across R&D worked very hard to create as leaving presents, the I&P team has used the magic of the previous synthetic voices project to create a Andy Bot reciting poems and be the MC of his own rd.io.