(Subscribe to our Today's Cache newsletter for a quick snapshot of top 5 tech stories. Click here to subscribe for free.)
Voice assistants use natural language processing (NLP) algorithm, an artificial intelligence (AI) model to train computers to understand spoken and written human language. The model better suits users accustomed to the technology as they know how to phrase their questions to the machine.
But older adults, especially those with limited experience in using latest gadgets, often struggle to ask conventional questions in a way the machine can understand.
To help senior users, a researcher at the University of California, San Diego, has devised a voice assistant that will focus on understanding and responding to medical queries from elderly users, the institute said in a statement.
The model is an end-to-end question answering system, and improves on the assistant’s ability to understand the question, instead of asking the user to restate it.
Also Read | A feature to make Google Assistant pronounce names accurately
Upon being asked a query, the system will break down the key components and match the shortened medical question to a frequently asked question from a database of 17,000 medical queries sourced from the National Institute of Health. Finally, it will select the relevant portions of the answer to share with the user, the researcher noted.
The researcher will train AI and NLP systems to summarise questions in a comprehensive manner. The model can be integrated into existing voice assistants like Alexa and Siri. The researcher also aims to make the system available for older and underserved communities around the world in native languages.
