This project will deliver a new set of machine learning (AI) models to classify COVID-19 human-based clinical studies with structured data from Cochrane's bioinformatics vocabularies. This innovation will enable the discovery, evaluation and synthesis of the most relevant and up-to-date COVID-19 primary evidence from around the world, to answer as quickly and comprehensively as possible specific clinical questions currently being prioritised by clinicians and healthcare policy advisers.

"Luminery is an innovative platform that harnesses crowdsourcing of media metadata to drive better revenue for publishers, and to create a new, fair and flexible crowd-working revenue stream for subject matter experts. Importantly, the crowd workers own the metadata they create - it is a data asset in this new marketplace. This metadata can be used by publishers a) to drive contextual commerce, b) to create brand moments, and c) to drive discovery and user experiences that lead to higher ecommerce conversion rates. During this proposed ""Phase 2"" project, we plan to explore these three, and other commercial opportunities, presented by the metadata. Unlike pure AI & machine learning -driven native commerce, Luminery harnesses experts' knowledge, via crowd-sourced tagging, to enable product discovery and better user experiences. There are opportunities to harness AI and machine learning to support this, and these will complement and enhance the human curation capability rather than replace it. The ""Phase 1"" market test, co-funded by Innovate UK and Data Language Ltd, is live and gathering data. This first market test uses MVP Luminery toolkits for crowd workers and video publishers, connected by the prototype feedback loop. This ""Phase 2"" will enable us to launch two additional market tests, to iterate based on the market test data, to explore the allocation of tagging work to crowd workers, to experiment on the feedback loop & quality assurance models, and to explore other user experiences that make up the commercial opportunities. Crucially, the experimental development will explore the key variables via a series of tests and iterative improvements, in real market settings, to optimise the core model and move the platform to commercial readiness."

A platform to enable the crowd sourced application of high quality and high value metadata to digital media to fundamentally transform the monetization of online video for publishers