Naver Cloud (CEO Kim Yu-won) announced on the nineteenth that it has accomplished constructing a small language model (sLM) specialized for the financial environment based on the lightweight model ‘HyperClova X Dash (HCX-DASH)’ in Mirae Asset Securities’ on-premise environment.
Naver Cloud said it has began a big language model (LLM) constructing business with major partners to enable public and financial corporations to utilize generative artificial intelligence (AI) in on-premise environments, and explained that the primary case is Mirae Asset Securities.
The build-type LLM is a model implemented inside the IT infrastructure owned by the corporate, not the general public cloud. Naver Cloud provided technical support, reminiscent of HCX-DASH and sophistication codes that may learn data owned by Mirae Asset Securities. Mirae Asset Securities explained that it built an SLM optimized for financial work by learning internal data.
Naver Cloud now has ▲ general cloud-based Hyperclova X service, ▲ Clova Studio, which allows customers to create customized AI services directly through the cloud, ▲ Neurocloud for Hyperclova X, which configures a network environment with a closed intranet inside the customer’s data center, and ▲ a ‘buildable’ service that could be installed and used on corporate infrastructure.
Tae-Geon Lim, Executive Director of Naver Cloud, said, “We’ve further strengthened our lineup in order that corporations can select and use it in the specified form depending on their usage purposes and environment,” and “Particularly, we expect that the build-type LLM will provide a possibility for public and financial corporations to utilize generative AI.”
Reporter Jang Se-min semim99@aitimes.com
