SEOUL -- Later this week, South Korea's largest web portal operator Naver will unveil the company's latest hyperscale AI solution in a bid to check global tech giants that develop large language model artificial intelligence-based solutions including Google and Microsoft.
A large language model (LLM) is basically a smart computer program that understands and generates human-like conversations in text. Through machine learning, an LLM is smart enough to engage in conversations and ask questions about different topics, and translate languages. An LLM is based on about one billion parameters, which are variables present in the model. OpenAI's ChatGPT, one of the world's most famous LLM-based chatbot services, is known to have some 100 billion parameters.
While it is easier for English language-based LLM developers to gain access to parameters for machine learning, it is harder for Korean language-based AI developers to obtain parameters for their Korean-based LLMs because there is a lesser number of datasets available for the isolated language with some 77 million Korean-speaking people across the world. A hyperscale AI solution refers to a system or infrastructure designed to handle massive amounts of data and computations for AI tasks on an extremely large scale.
According to Naver, the tech giant will unveil "HyperCLOVA X," a hyperscale LLM, on August 24. The new LLM is a brainchild of the South Korean counterpart of Google which released "HyperCLOVA," utilizing data and a supercomputer hooked onto Naver's service ecosystem, in 2021. Naver's new LLM has some 204 billion parameters.
The number of HyperCLOVA X's parameters is larger than "GPT-3" (175 billion parameters) and has studied Korean language-related corpus by about 6,500 times more than ChatGPT based on the "GPT-3.5" system. During the showcase, Naver would introduce blueprints for business-to-business (B2B) services that would connect HyperCLOVA X to various businesses in the commerce, creative, finance, and other sectors.
reporter
Park Sae-jin
swatchsjp@ajunews.com
