East Lansing Company Develops Sentiment Analysis Tool for AI
East Lansing’s Virtual Emotion Resource Network (VERN) has patented a sentiment analysis tool that helps artificial intelligence systems detect human emotions in messaging.
The company and its founder, Craig Tucker, hope to someday develop computerized systems that can enhance customer capabilities of sentiment analysis across public and private sectors touching every industry.
While this technology, which the developer says outperforms existing AI technology, has many potential uses, increasing mobility in email communication and enhancing non-player characters (NPCs) in gaming environments are early targets.
According to Tucker, VERN features a “revolutionary” new model of communications, an incongruity detector, and a machine learning capability to refine the preliminary scores into a final emotive score. Available through an API, it allows companies that have AI systems to use VERN to analyze the emotional content within their messages.
“Our competitors offer users only positive, negative, neutral, or mixed over an entire message,” Tucker says. “VERN analyzes each sentence for emotion and gives the emotion over the entire message.”
The humor detector will be produced first to fill a void in that there are no other commercially available humor detection algorithms, according to Tucker. “We are on pace and in the coming months, VERN’s additional emotives and machine learned analytics will eclipse the effectiveness of our client/competitors.”
Tucker currently is in the Ph.D. program at Michigan State University. VERN is in the process of raising funds to continue development and bring the emotion-detecting tool to market.
VERN expects to generate $3 million in revenue in 2021 and grow to $8 million in 2023 with a net Income of $1.1 million and $6.2 million respectively.