2nd September 2019
Civica enlightens library search with artificial intelligence
Leverages Microsoft technology developed for JFK data
Civica has built a cloud-based AI infused library search and management platform, leveraging technology first developed by Microsoft to search and interpret all the documents associated with the death of President Kennedy.
A global software company focussed on solutions primarily for the public sector, Civica has enhanced its Spydus Library Management System so that it can ingest a wide variety of content, including photographs, documents, newspaper clippings and letters – then using Microsoft Azure Cognitive Services make that easily searchable and accessible.
The underlying technology was used to similar effect by Microsoft to analyse the trove of documents which the US Government has released in recent years, regarding the 1963 assassination of JFK.
Now Civica’s market-leading Spydus, already widely used to manage the day-to-day operation of more than 2,000 public library services in Australia, New Zealand, Singapore, US and the UK, has been transformed by AI.
According to Simon Jones, Managing Director Libraries & Educations solutions for ANZ, Civica; “Libraries are increasingly taking on the responsibility of being a community information repository – a home for all manner of information, documents, letters, photographs, video and audio files that make up our oral history. The challenge is to make this rich data set available for further use.
“However, libraries are already busy places – they don’t have the time or staff resources to manually catalogue, index and manage all of these data resources. Using Microsoft Azure Cognitive Services we have been able to automate the task – and in the process also enhance the quality of meta information making these data collections more accessible, and hence more valuable.”
To the community this democratised access to information is invaluable. It supports people researching their family trees or community heritage groups – making information that was previously hard to access much more discoverable.
The digitised and AI-infused solution also increases the accessibility of information for hearing or vision impaired users who can benefit from the accessibility and access tools that are available in the Azure platform.
The reach and scalability of Microsoft’s Azure platform means that access is available regardless of location – an important factor for historians or researchers who want to access information remotely. Meanwhile library users will be able to more easily search content for their own investigations into family or local history, for example.
Leveraging the same sort of technology used in Microsoft’s JFK proof of concept, Civica, working with Stonnington Libraries in Victoria, has built a prototype solution. It loaded a large quantity and variety of data from Stonnington’s collection into Azure Blob Storage, then used Azure Cognitive Services to create meta data that would allow the repository to be searched.
In the proof of concept with Civica has focussed on two media types; digitised images such as photos, maps, scans and sketches as well as long form text style documents including minutes of meetings, flyers, newspapers, articles and community group newsletters.
Looking forward it wants to include audio files to allow oral histories to be accessible as well as contemporary video content that will be of interest to future generations. Azure supports the storage and access of these content types already, so the platform is future proofed to allow additional types of content capture and access over time.
Rachel Bondi, General Manager, One Commercial Partner, Microsoft Australia said; “AI is transforming all our lives – this is a perfect example of how AI can democratise access to information that was previously out of reach.
“Civica’s innovation means that the rich troves of community-supplied information can now be stored securely, categorised automatically using Microsoft Cognitive Services, and made available for analysis through the cloud ensuring universal access. We are delighted to be working with Civica on this important initiative.”
The prototype developed in association with Stonnington Libraries used its existing archive database that has been manually catalogued over the last 20 years. Civica took the raw digital files, photographs and PDFs and fed them into the Azure data collection.
Using Microsoft AI it was able to create a searchable catalogue in just a couple of hours.
In order for the content to be made publicly searchable the owner or donor of the information assigns copyright for that purpose, though that will not always extend to re-use of the materials.
Libraries may, for example, use the platform to create new services and products beyond simple search. A library may choose to make some images available publicly, but sell high resolution images for reuse. Azure AI resources can also be used to help manage the use of content.
Civica’s platform is being developed to support a wide range of use cases – but individual libraries will determine how they use it.
Natasha Tsui-Po, Senior Team Leader, Systems & Resources, Library & Information services at Stonnington Library; “You cannot underestimate the value of local community information – but creating the access to it has always presented challenges as each item has to be manually indexed, described and categorised. With Civica’s new approach to automate that initial task, these resources can now be accessible much faster and without the burden on already time-pressed library personnel.”
The AI infused approach also supports more comprehensive data tagging than human librarians typically perform, increasingly the likelihood that information would be surfaced during an end user search. Civica expects that the greatest impact will however come from the combination of AI and human effort – bringing together the speed and accuracy of AI, enhanced by the nuance of human insight.
Civica is extending its development, exploring the opportunity to use voice to text technologies that would make audio and video materials even more discoverable.
It also envisages the platform being used to underpin entire new classes of library services such as community portals allowing online exploration of information collections.