Helping staff to find relevant learning resources through intelligent search and automatic semantic annotation
Senior Business Developer
Maintaining a skilled and knowledgeable workforce for all National Health Service (NHS) staff is crucial. The NHS is the world’s 7th largest employer with over one million+ doctors, nurses and ancillary staff. The COVID-19 pandemic only exacerbated the need for easy, intuitive access to the best training resources available.
There was a need for creating a new NHS Learning Hub where staff can quickly find the most suitable and relevant training, educational and course resources. A Hub that can unify resource access for the tens of thousands of resources that lie across hundreds of learning portals. Heterogeneous resources also come in all types of media and presentation formats: sessions, courses, programmes and catalogues.
The growth in healthcare knowledge is second to none. This means constant additions and versioning of resources in order to reflect new knowledge. Resources had poor and limited metadata with inconsistent editor tagging (folksonomy), meaning that resources remained hidden.
The solution includes i3 (intelligent information integration), PoolParty (knowledge engineering tool), Elasticsearch (open source search engine) and Kibana (an open source data visualization dashboard for query and click log data).
i3 processes enriches resources before indexing. As part of the process, concepts extracted from the resource metadata (title, description etc.) are referred to the Health Education England (HEE) healthcare knowledge graph, in order also to find how they relate to other concepts.
In doing so the “aboutness” of a resource is calculated and it becomes annotated with the most appropriate keywords – making them “smarter,” and therefore more findable. This includes annotations where a keyword concept may not actually appear in the resource metadata (a shadow concept) by default of other the concepts and their relationships that appear in the knowledge Graph.
The Learning Hub solution went live in May 2020, six months earlier than scheduled due to COVID-19.
The addition of structure, meaning and context to learning resources has resulted in both intelligent search and a more intuitive GUI (fed by Findwise i3 APIs to HEE GUI). Search and the knowledge graph are continuously improved through the monitoring and analysis of end-user signals and behaviours.
There are many new exciting features and additional functionality currently being developed with user research feedback that will be released as part of the agile Learning Hub roadmap.