This could lead to the next big breakthrough in common sense AI

This could lead to the next big breakthrough in common sense AI


Publication Title
MIT Technology Review
Publication/Creation Date
November 6 2020
Creators/Contributors
Karen Hao (creator)
University Of North Carolina (contributor)
Hao Tan (contributor)
Mohit Bansal (contributor)
Persuasive Intent
Academic
Description
"You’ve probably heard us say this countless times: GPT-3, the gargantuan AI that spews uncannily human-like language, is a marvel. It’s also largely a mirage. You can tell with a simple trick: Ask it the color of sheep, and it will suggest “black” as often as “white”—reflecting the phrase “black sheep” in our vernacular.

That’s the problem with language models: because they’re only trained on text, they lack common sense. Now researchers from the University of North Carolina, Chapel Hill, have designed a new technique to change that. They call it “vokenization,” and it gives language models like GPT-3 the ability to “see.”"

HCI Platform
Other
Location on Body
Not On The Body
Source
https://www.technologyreview.com/2020/11/06/1011726/ai-natural-language-processing-computer-vision/

Date archived
November 9 2020
Last edited
November 10 2020
How to cite this entry
Karen Hao. (November 6 2020). "This could lead to the next big breakthrough in common sense AI". MIT Technology Review. Massachusetts Institute Of Technology (MIT). Fabric of Digital Life. https://fabricofdigitallife.com/Detail/objects/4994