Galen has been working on his first two games designed to give 12-15 year olds good intuitions about aspects of machine learning. He's currently getting some feedback within R&D on his first one, AI Artschool (see screenshot below); David's been helping him with the design and the feedback form.
David has also been working on designs for our Images of AI website, and Alicia has been implementing them. This is a site for better stock photos of AI. Henry's busy working with an artist to create a brief for artists to contribute (paid) work to the site. We've also been working on a microsite with for their students' responses to the brief we gave them on this topic.
Tristan and Kristine also presented our work at the R&D allstaff meeting.
In non-machine learning news, Henry and Libby have been designing and practicing another futures workshop, building on our signals gathering and impact wheel workshops, this time about creating scenarios illustrated using mockups of mundane things you might find in the future.
Meanwhile, the Data team have been taking part in a corporation-wide hack week that was organised in partnership with Google. It was built around the themes of education, diversity and inclusion, attracting new audiences and creating higher impact content. The event was was a chance for us to meet and work with staff in other parts of the business and prototype some new ideas. Andy worked on an idea to draw users into archive journeys using clever interface designs, Ben worked with his team on processing and extracting metadata from archive content using NLP, Rob and Ollie worked together on an prototype involving automated idiom detection and Polina was the lead researcher on an idea for an improved bandit recommender.
In other Data team news, Ben has been looking at the comparative performance of running iRex algorithms in various flavour of Python, Fern is about to start some important work looking at Audioset embeddings to discriminate subgenres in programmes and Kristine is looking at applying a new tagging system for ±«Óãtv Sounds as part of her ‘Introducing..’ work. Finally, Matt is doing great work sorting out various infrastructure issues: looking at the troublesome power supply for our GPUs, tidying up various S3 buckets and planning the final decommissioning of Snippets.
The Interaction & Prototyping team have also been busy. They have made a first draft of their social interaction toolkit document - the design patterns and toolit developed from their multi-person game prototype. They spent a day laying out ideas and opportunities for new prototypes that would showcase the toolkit's benefits, which created some good discussions about potential additions to the existing patterns. Andrew and Mathieu have drafted the first blog post about the project, chronicling its initial aim and outlining the features of the prototype.
Barbara has been running a workshop around remote connected cars with the UX&D NeX team as part of the Innovation Working Group. Colleagues from UX&D, Voice, Sounds, Children, TS&A, R&D North and R&D South labs got together in groups to flesh out ideas for new experiences the ±«Óãtv can enable in-cars. Responding to opportunities and user needs identified in M&A’s In-Car Audience research, the groups produced self-contained demos that can be used as vision pieces to convey the value to the user. The feedback was excellent (despite the heat wave), with great engagement and some really innovative ideas coming out.