Published : Feb. 25, 2021 - 15:35
LG Twin Tower (Yonhap)
LG Group unveiled its first achievement in artificial intelligence research conducted by the group’s AI research arm.
The South Korean conglomerate announced two academic papers on explainable AI and continual learning in collaboration with University of Toronto, through the Association for the Advancement of Artificial Intelligence.
Explainable AI, co-researched by a team led by Professor Konstantinos Plataniotis, refers to a technology that easily explains a result of its research to humans by providing reasons and evidence.
For example, when the explainable AI is applied to analyzing X-ray images, the AI system not only provides a result, but also how it has come to the conclusion.
This could be used to support or replace human decision-making in professional fields like medicine, finance and law.
“The explainable AI is a nascent sector, and even Google is in very early stages of research on it,” said LG.
Continual learning enables AI to continue its learning data sequentially and to make a comprehensive judge by analyzing the data accumulated from past learning.
LG and a team led by Professor Scott Sanner have improved the learning capability by 40 percent by applying the Sharpley value – a measure for importance of data – for the first time in the field.
“LG aims to make our customers feel and experience AI technologies by elevating the levels,” said Bae Kyung-hoon, head of the LG AI Research. “Our goal is to nurture 1,000 AI experts by 2023.”
By Song Su-hyun (
song@heraldcorp.com)