FutureFive New Zealand - Consumer technology news & reviews from the future
Story image
Virtual assistants could soon help those with mental illness
Wed, 18th Sep 2019
FYI, this story is more than a year old

 Computer scientists have created a new way of using virtual assistants that could potentially help those with mental illnesses.

Researchers at the University of Waterloo created SMERTI, which is a method that allows virtual assistants to use natural language and emotional cues that change based on the way they are used.

Researchers hope that the end result will help to develop virtual assistants that connect better with the people they're designed to help.

“Certain personalities or emotions within a virtual assistant appeal more to certain individuals,” explains Waterloo undergraduate student Steven Feng.

“Enabling virtual assistants, based on the situation, to tweak the words or sentences used to match the personalities could lead to on-demand chatbots available to talk to people with mental illness and cognitive disabilities whenever they require.

He adds that more and more people are requiring mental health support, but there just aren't enough therapists to help.

“Being able to automate some aspects of treatment will be beneficial as it would reduce wait time and make the process more affordable. But the emotional aspects of mental health are a major challenge for virtual assistants.

SMERTI is one of several artificial intelligence-based tools that use similarity masking (SM), entity replacement (ER) and text infilling (TI).

It is able to take a virtual assistant's personality and adjust text responses based on current situations.

 For example, it would take the advice of “It is sunny outside; I know you hate to, but you must wear sunscreen”, to “It is rainy outside; I know you hate to, but you must bring an umbrella”.

Researchers tested SMERTI by presenting eight other researchers with original text written by humans, in addition to modifications from multiple tools, including SMERTI.

The participants were then asked to blind review the sentences and rate them on a scale of one to five based on three factors: Fluency, sentiment preservation, and content exchange. Researcher could then see how well the new text worked in place of the human-written text.

“Based on the respondents' ratings of the various systems it was found that SMERTI outperformed all the baseline models especially in terms of fluency and overall replacement of the text to fit the new semantics,” says Waterloo Cheriton School of Computer Science associate professor Jesse Hoey.

“What we were mainly focused on was the fluency and semantic exchange aspects to show that the task is possible and measure the emotional preservation, which was decently high. The next step is to research how to preserve personality or persona rather than just emotion, which is more complicated.

The researchers are now uncovering how they can develop text that matches a virtual assistant's personality.