Themes: Assistive & Accessible Technology
Improving information retrieval about assistive technology through an intelligent insights portal
Policymakers and decision makers need access to good quality information about assistive technology (AT) in order to set policies and strategies in this field. Yet information about AT is often found in disparate locations, and it’s not always easy to find. PhD student Sahan Bulathwela is developing an intelligent insights portal to make it easier to retrieve key information about AT.
In order to understand and plan for the AT needs of populations, policymakers and strategic decision makers need to access information about AT to create reports, white papers and policy documents. While the internet is full of information, sifting through what is relevant and high quality is difficult. In addition, information about AT is often held by multiple different organisations, and there is no single place to identify all relevant information about AT.
As such, there is a need for a centralised system that can pull together relevant information about AT.
Using technology to manage AT information
In response to this, PhD student Sahan Bulathwela is developing an intelligent insights portal.
“I am using artificial intelligence (AI) to match the right information to the right person at the right time, this portal is being developed so policymakers can take more informed, data backed decisions in their work. For example, if a person in the World Bank wants to build a strategy about how they want to spend money over the next 10 years, how do they gather the right information to make those decisions? The system I’m building is designed to help them with that.”
To start with, Bulathwela is identifying the different types of information that exist on AT. This includes media feeds, publications and reports from a variety of different sources. He then plans to process this information using AI, annotate it, and build content representations to cater to the information needs of policymakers and decision makers.
Bulathwela has also identified the different systems that are currently available to facilitate complex data retrieval. He is exploring how AI can be specifically used to inform users about collections of documents that are relevant to them. He has been running some of these state of the art systems, and is currently honing the technology to look specifically at retrieving data about AT.
Next steps
Once the technical processes are ready, Bulathwela will run some user studies to understand if the system meets the needs of policymakers and decision makers interested in AT. He will then build the system, and run more rigorous user studies to refine and validate the portal. Eventually, Bulathwela would like the system to be used by a range of users, including those with AT needs themselves.
Bulathwela’s PhD is on building educational recommended systems using AI, and this intelligent insights portal is a use case for this wider topic.
“The insights portal is trying to address the personalised information needs of different stakeholders,” Bulathwela said. “In terms of policymakers and decision makers, I envisage most will use this system for compiling reports. For example, if someone wants to know what is happening in terms of wheelchairs in the world now, they’ll be able to see what the media is saying about it, what the trends are, what organisations are being mentioned in relation to wheelchairs, and what scientific publications and grey literature exists on it. This will help them decide where to allocate money or where they should be calling for proposals.”
Funded by: AT2030
