Computer Scientists have developed a new method that could be used to develop more” natural and human-like” automated virtual assistants that will help people suffering from mental illness.
The new system, called SMERTI (pronounced: “Smarty”), enables the virtual assistant (VA) to use natural language and emotional clues that vary depending on the relationship and situations they are used in. The result will allow the development of virtual assistants that connect better with people they are used to help.
The number of people who require mental health therapy is increasing day-by-day, but there are not enough therapists to deal with these cases. Automating some aspects of treatment will be beneficial as it would reduce waiting time, overlapping appointments, and make the process more affordable. The emotional aspects of mental health are a major challenge to overcome for virtual assistants.
Steven Feng, an undergraduate student in Waterloo’s David R. Cheriton School of Computer Science explained that certain personalities and emotions in a virtual assistant appeal more to individuals. Enabling virtual assistants, based on the situation, to tweak 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 are required.
SMERTI represents some artificial intelligence software tools working together, including similarity masking (SM), entity replacement (ER) and text infilling (TI). SMERTI takes a text response from a virtual assistant that has a certain type of personality and adjust it so that it fits the current situation. For instance, it will 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.”
To evaluate the system, the researchers presented eight fellow researchers with original pieces of text written by humans, with various modifications from multiple tools, including SMERTI. The respondents were asked to review the sentences and rate them on a scale of one to five based on fluency, sentiment preservation, and content exchange.
Based on the ratings of the various systems, it was found that SMERTI outperformed all the baseline models in terms of fluency and overall replacement of text to fit the new semantics, said Jesse Hoey, an associate professor in the Waterloo’s Cheriton School of Computer Science.
“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.” said the team.
The researchers are now working on a system to generate the personality-favored text, which, when combined with SMERTI, will result in virtual assistants that have a more consistent personality as consistent personality is a key component of future virtual assistants for mental health therapy and care.