Metabuild Expands sLLM Business Area for Ultra-Large Natural Language Model 'LLaMON'
Metabuild, an AI company, said, "2024
will be the first year of its second growth" with its ultra-large natural
language foundation model, "LLaMON." Along with its End to End (E2E)
data fabric product line, it is strengthening its AI LLM/sLLM application
service business such as "Document creation, summary, consultation,"
"AI music composition service KEENEAT," and "Digital Human
Service (R2MIX)."
In its 25th year, Metabuild is supplying
'MLOps-based AI Platform (MAI AUTO), 'MESIM ESB/APIG', 'MESIM DATAHUB',
'Network-based SW (MESIM mHUB)' and 'Digital Twin-based Integrated Control
Platform (Transportation, Autonomous Vehicles, City, Construction, and Drone)'
to 4,000 institutions in various fields such as administration, industrial
complex, manufacturing, communication, city, road, and construction.
'MESIM ESB/APIG' provides operations in a
hybrid manner of on-premises and SaaS/gSaaS cloud environments. It is a product
that connects and integrates large amounts of data between various and complex
internal and external big data platforms, MyData platforms, and digital
platforms, and it is possible to easily develop linked services with template
development tools provided with various real-time linked agents such as open
APIs. It provides independent interoperability in the system environment such
as large-capacity processing, flow control, high availability, and integrated
monitoring. 'MESIM mHUB' dramatically reduced the delay time in processing
real-time traffic between dual network, thereby increasing operational
efficiency and is the highest performance and quality product selected as the
No. 1 in the TTA's open BMT. With the advent of an aging population of 10
million, chronic patients with industrial accidents, mobility discomfort,
diabetes and high blood pressure can receive comprehensive medical
rehabilitation delivery reservation services, diabetes management services,
medical device long-term care services, and simple administrative services by
utilizing Metabuild's 'PILLLUCK'.
'MAIAUTO' supports the entire process of
developing AI services, from data collection to design, learning, distribution,
and management of AI models. By applying its 'MAI AUTO Ops AI Development Methodology',
it has excellent AI innovation cases in the medical field by establishing an
'Unfair detection and prediction' service at the Korea Natonal Health Insurance
Service and the Health Insurance Review and Assessment Service. 'MAIAUTO'
provides both data-based and ultra-large natural language-based. Both beginners
and experts can easily develop advanced AI technologies such as data
collection, storage, analysis, and processing of big data for the development
of various AI services, based on UI/UX tools. It is possible to easily
configure AI services by providing data processing and AI development APIs in
the cloud and on-premise environments. Non-experts with no knowledge of AI
development can also develop by a UI that allows users to easily use the entire
process from data preprocessing to AI model development. It is characterized by
improving the existing method that users develop themselves, which can reduce
work time and increase efficiency.
"LLaMON" provides LLM models such
as Lamar 2, GPT 3.5, and Palm 2 or sLLM models such as BERT, BART, and GPT
depending on the AI service operation environment, and provides a "LLaMON
LLMOps platform" to efficiently support the introduction and construction
of natural language processing AI.
LLaMON provides various sLLM models used to
build natural language processing services for general consumers in a cloud
environment. Since data cannot be leaked outside depending on the
characteristics of public institutions or companies, it supports sLLM to create
language models specialized for companies by learning data held by companies in
the form of on-premises, where it is difficult to invest large-scale resources
such as cost and time.
The LLaMON LLMOps platform is characterized
by the ability to implement language processing solutions that meet the
specific needs and conditions of the company. It can be used in various
services by providing ultra-large language models to analyze and model various
natural language processing problems required by AI services that companies
want to develop.
"Digital Human Service (R2MIX)"
equipped with "LLaMON" can coordinate the skin tone, hairstyle,
costumes, gestures, and background of a digital human, and follow the
expressions of media creators on a smartphone or laptop equipped with a camera,
or synchronize and broadcast in real time when a script is given.
The "KEENEAT AI Music Composition
Service," which was learned with "LLaMON," provides high-quality
music of my taste with the cloud so that even non-professional ordinary people
can easily and quickly compose music.
Metabuild will present cutting-edge products such as "LLaMON," "Digital Human Service," and "Digital Twin Control Platform for Traffic, Autonomous Vehicles, Cities, and Construction Services" at the world's largest IT and home appliance exhibition "CES 2024," which will open in Las Vegas on January 9, 2024 (local time). With "CES 2024," Metabuild plans to strengthen its entry into overseas markets of its flagship products such as "LLaMON" and "AI Music Composition Service KEENEAT," which are recognized as the best technology products.
Published: 26 Dec, 2023