Search SciPol

Brought to you by
March 7, 2017

IBM Watson Takes On Autism

  • Other
  • Neuroscience
  • Robotics/AI

Forbes – IBM Watson burst onto the world stage in 2011 when it participated in the trivia-based game show Jeopardy!. The supercomputer beat out two former champions to claim a victory for “artificial intelligence”. Since then, Watson has embarked on a number of challenges across a variety of domains, from identifying the best cancer treatments to improving weather forecasting. For its latest endeavor, Watson is looking to improve the quality of life for individuals with autism and other cognitive disorders.

Autism refers to a group of complex disorders of brain development characterized by difficulties in social interaction, verbal and nonverbal communication and possible repetitive behaviors. The current incidence of autism is 1 in 68 children.

When the media reports on stories related to supercomputers and autism, the topic is usually centered around an attempt to find an “autism cure” through better understanding of genetics and biomarkers. However, what makes this effort interesting is that it’s focused on making life easier for individuals who already have the disorder. Projects such as these are part of a wider societal movement for progressing beyond “awareness” for individuals with special needs, and towards “inclusion and acceptance”. As the father of a wonderful young man with autism, I can certainly appreciate the importance of such efforts.

The new Watson application, called Content Clarifier, was developed on Bluemix, IBM’s cloud platform, by IBM Accessibility Research and represents a powerful combination of machine learning, cognitive technologies and natural language processing. The system takes a long and complicated body of source content and dynamically filters out unnecessary information, replacing complex and recurring words or phrases, and converting it into something more understandable for the user. The content is essentially tailored to their specific needs.

Dr. Will Scott, IBM’s Lead Software Architect on the project, explains further. “Using a cloud-based infrastructure and cognitive computing, Content Clarifier analyzes and condenses content into clearer language, such as by substituting euphemisms with more straightforward verbiage, by adding visuals, or by inserting contextually relevant semantic content (i.e., maps, web resources, phonetic pronunciations). It greatly simplifies the task of reading with the end goal being to increase overall comprehension.”

By delivering highly customized and personalized content (e.g., plain language definitions, alternative augmentative communication symbols, or speech), the Watson and IBM cloud-based system can improve the quality of life for people with a variety of cognitive challenges.

Scott explains that "individuals on the autism spectrum can benefit from the technology in a number of ways. They may find learning to read very challenging, or may read at an early age but without understanding the meaning (a syndrome identified as hyperlexia)." Content aggregation and simplification can certainly help address both of these challenges. Additionally, Content Clarifier is being extended to utilize Watson’s Tone Analyzer to better understand the emotional tone and style of written text. This further increases a user’s comprehension of complex content.

Scott added that “… ongoing collaboration with subject matter experts in academia is key, and we will continue to feed findings from this research into our development process to iterate on and improve the system.” One example is IBM’s collaboration with the University of Massachusetts Boston. Their task is to research the effectiveness of the Content Clarifier system against the output of human translators. This comparison helps to understand how to continuously improve natural translation, simplification and summarization capabilities.

Content Clarifier has a number of other uses beyond helping individuals with autism. It can also be used for a variety of applications, such as helping those learning English as a second language better comprehend job postings and other digital content, or assisting elderly individuals with cognitive decline to better understand the world around them.

Once people are able to automatically personalize complex information into a simpler format, the world becomes a little less intimidating … for everyone.

Image citation: Rosemaryetoufee, CC BY-SA 4.0