LG AI center joins forces with Parsons School for creative AI development 

South Korea’s LG Corp. announced it signed an academic collaboration agreement with the New York City-based Parsons School of Design on Monday.

The LG AI Research plans to work with the world-renowned institute to further develop its artificial intelligence (AI) to be used in the design and fine arts sectors. 

The two sides formed a four-year LG-Parsons Creative AI Research Partnership.

The AI institute has developed a hyperscale AI called Exaone, short for EXpert AI for everyONE, widely thought to be capable of learning and making decisions like the human brain. 

LG AI Research and Parsons will host a symposium and a hackathon this year, as part of the joint effort to establish the research structure for the next three years. 

The full-scale joint study will begin next year, which will focus on developing generative AI technologies and methodologies that can be used in design and arts – utilizing LG’s hyperscale AI system Exaone. 

The research center plans to use the findings from the joint study to advance its AI service for collaboration with designers and artists.

AI ATELIER PLATFORM

Last month, LG AI Research launched Exaone Atelier, a creative platform in which a designer can collaborate with AI to generate new visuals based on prompts. 

The new platform can generate 256 high-resolution images in seven minutes when given a single sentence. 

For instance, when asked to come up with a fox character for a nursery tale, the AI generates several different versions. 

“Through our collaboration with Parsons, we will create an AI designer that can maximize human creativity and productivity,” chief of LG AI Research Bae Kyung-hoon told The Korea Economic Daily. 

In February an LG-developed AI designer named Tilda debuted at New York Fashion Week in collaboration with a famous South Korean fashion label.

According to LG, Exaone boasts a whopping 300 billion parameters and is equipped with multi-modal capabilities to acquire and process information related to nearly all aspects of human communication, not just written and spoken languages.

The research institute has trained Exaone to study 600 billion text corpus and 250 million images simultaneously. 

Parameters refer to where AI’s learned data via deep learning gets stored and is often used as a measure of how well a model is performing. In human physiology, they are somewhat comparable to synapses. 

Write to Ji-Eun Jeong at jeong@hankyung.com
Jee Abbey Lee edited this article.