PRE™ Uses AI Bias for Good with MultiVersal AI Demographic Simulator
TORONTO, May 18, 2023 (GLOBE NEWSWIRE) — PRE, a Toronto startup that is revolutionizing e-commerce sites and apps using AI, announces the launch of its new MultiVersal training system. This new AI program pre-filters training data to highly specific demographic segments and updates the human behavior model parameters based on the filtered data. Subsequent applications of PRE give unparalleled specificity for e-commerce testing and optimization, giving brands the tools to create the most optimal online retail experience possible for their particular user demographic.
“A key request from our partnered brands has been for the ability to focus our AI on a particular demographic. With MultiVersal training, we can do precisely that,” said PRE Founder Parham Aarabi. “We are creating bias to analyze the demographic variability of AI systems, enabling partners with an unprecedented capability for optimizing their e-commerce sites and apps.”
The MultiVersal system creates AI bias to test its impact on the simulated e-commerce site performance. For example, companies can use it to bias the training data toward older female users to measure how the site performs for that particular demographic. PRE’s MultiVersal system can then change the focus to a different user segment (e.g. middle-aged women, younger men, etc.), redo the simulation, and obtain results focused on the new user segment in real-time.
Nearly all existing AI systems today exhibit some degree of bias due to training data limitations. These systems are trained on “universal” demographic data, which only allows them to perform at a holistic level without the ability to model a specific user segment. MultiVersal training fixes this problem by enabling sites to dynamically match PRE’s behavior model to a specific user demographic.
For more information about PRE and its capabilities, visit https://pre.cm. Contact [email protected] to set up a personalized demo.
PRE is a leading AI platform that simulates e-commerce sites/apps using AI-generated virtual shoppers. Founded by Dr. Parham Aarabi, whose previous work includes founding AI/AR giant ModiFace, PRE simulates e-commerce sites 1,200 times faster and 100 times more cost-effectively than current options, all the while maintaining an accuracy of 94%. For more information on how to incorporate PRE for your e-commerce sites, visit https://pre.cm.