Subscribe
E-Paper
 |
| LG AI Research Center at LG Science Park in southwestern Seoul / Courtesy of LG |
By Kim Jae-heun
LG's artificial intelligence (AI) researchers are moving quickly to lead the convergence trend of visual AI and language AI.
The company's research center will hold the LG Global AI Challenge for AI researchers around the world from Feb. 1 until the end of April, LG said Tuesday.
To hold the competition, LG will establish a joint research center with Seoul National University's AI Graduate School and U.S. stock image provider Shutterstock. They are developing and preparing for the commercialization of AI EXAONE, an image-captioning AI service.
"Through this competition, we would like to create a place to discuss the significance, necessity and expandability of AI research with experts around the world," an LG AI researcher said.
LG Global AI Challenge will be held online to evaluate how accurately an AI understands and explains images presented to it, under the theme of "zero-shot image captioning."
Zero-shot image captioning is a technology that enables AI to explain the results of its understanding of an image that contains an object, animal and landscape based on data accumulated beforehand.
It works similarly to a human's image recognition ability by distinguishing the different characteristics of certain objects.
As zero-shot image captioning advances over time and an AI's image recognition accuracy improves, LG believes the technology will have real-life applications.
