Superb AI (CEO Kim Hyun-soo) announced that it is going to expand its current platform, centered on vision AI, to a 'multimodal' basis to satisfy the increasing demand for artificial intelligence (AI) from firms. Through this, the intention is to expand business areas centered on safety and security.
A Superb AI official said on the twenty fourth, “We plan to put in multimodal vision AI on CCTV throughout the first half of the yr to enable natural language-based video data search.”
This can be a feature that was recently announced on the Mobile World Congress (MWC). Since its inception, Super AI has focused on introducing vision AI using images, videos, and 3D LiDAR. Subsequently, unlike other firms based on large language models (LLM), it’s a format that attaches a language model backwards to the prevailing Vision AI.
Through this, functions which were lacking in vision AI to date might be realized. For instance, by explaining raw data corresponding to CCTV in natural language, the user can quickly find the relevant part when searching in the shape of a sentence or word. When utilized in criminal investigations, etc., inputting the suspect's appearance and clothing might help with quick information retrieval.
Vision AI has been leading the AI industry within the manufacturing industry, taking responsibility for factory safety and defective product identification. Specifically, interest in overall AI has increased significantly as a consequence of the prominence of generative AI last yr, and up to date adoption has been on the rise. In this case, SuperbAI's intention is to expand the market with multi-modal services.
This company is considered one of the leading players in 'industrial AI', having developed greater than 3,000 customized vision AIs. Already within the manufacturing and steel fields, AI solutions have been introduced for product quality inspection and recycling sorting, reducing costs and time in addition to securing consistent quality standards.
Because it has quite a lot of cases and uses, Superb AI said it’s quickly feeling the recent increase in demand.
An organization official said, “Just just a few months ago, there have been many cases where the importance of terms corresponding to MLOps and data weren’t recognized, but now the demand has increased significantly, with more firms recognizing the difference in data pipelines in addition to terms.”
Specifically, it’s an organization recognized as having world-class technology in data labeling. “For that reason, we now have recently received inquiries for the aim of manufacturing educational and simulation videos.” The reason is that the platform shows the info labeling process in an easy-to-understand manner.
The all-in-one platform launched last yr is a culmination of Superb AI know-how and a product that became the stepping stone for a full-scale business. It supports the development and operation of a whole pipeline that repeats all the AI development cycle, from ‘data construction-selection-processing-management-analysis’ to model learning and distribution.
An official said, “We’re feeling a definite increase in demand since constructing the all-in-one platform. Specifically, there was quite a lot of demand for technology from practitioners previously, but recently, there are numerous cases where executives themselves are stepping forward.”
It is claimed that the rise in demand occurs in two ways: ▲domain expansion ▲overseas demand.
In the primary case, beyond the prevailing manufacturing or process areas corresponding to 'machine vision-based defective product selection', demand in the security and security areas is now expanding. A representative example is checking whether safety helmets are worn and protected distances are maintained within the factory. Demand is steadily increasing in places where data security issues don’t occur, corresponding to road CCTV.
Japan, a producing powerhouse, has also emerged as a serious market. He said that he’s feeling the demand after establishing a Japanese corporation following the U.S. corporation.
Accordingly, Super AI recently began collaborating with Rebel and is pursuing 'constructing a vision AI model based on closed network and edge AI'. It is because the manufacturing industry is reluctant to make use of cloud-based software as a service (SaaS) for security reasons.
Meanwhile, Superb AI is scheduled to take part in the '2024 Smart Factory-Automation Industry Exhibition (SFAW 2024)' held at COEX, Seoul from the twenty seventh to the twenty ninth. This exhibition provides product demos in addition to the chance to experience an AI-applied smart factory using security, CCTV, and 3D LiDAR.
Kim Hyun-soo, CEO of SuperbAI, said, “Interest in AI technology has increased even in traditional industries corresponding to manufacturing, but there still appear to be many cases where there may be a vague sense of apply it to industrial sites and use it.” He added, “AI is utilized in actual industrial sites. “We plan to point out through a demo apply it,” he said.
Reporter Jang Se-min semim99@aitimes.com