Lenovo Korea and Lomin Launch AI Solution for Personal Data De-Identification
IT DAILY ·
✦ AI Summary
Lenovo Korea and Lomin announced on the 20th that they had launched the integrated solution "Lomin Redact AI Appliance."
The solution consists of ThinkSystem servers, NVIDIA GPU-based computing infrastructure, and Lomin's de-identification software, "Lomin Redact."
Because detection and de-identification tasks can be completed within an in-house standalone infrastructure, sensitive original data does not need to be transmitted externally.
Lenovo Korea and Lomin announced on the 20th that they had launched an integrated solution, "Lomin Redact AI Appliance." The solution combines AI infrastructure with de-identification software.
The two companies pursued this collaboration to help businesses and institutions that handle sensitive data, including personal information, meet regulatory requirements and maintain data security. To that end, they adopted an approach that provides hardware infrastructure and specialized de-identification technology in a single environment.
"Lomin Redact AI Appliance" comes in an all-in-one appliance format and consists of "ThinkSystem" servers, NVIDIA GPU-based computing infrastructure, Lomin's de-identification software "Lomin Redact," and an operating environment.
The solution is structured so that sensitive original data does not need to be transmitted externally. Detection and de-identification tasks can be completed within an in-house standalone infrastructure. As a result, infrastructure deployment complexity is reduced, and security risks are also lowered.
Lomin Redact is embedded as a software engine and applies VLM technology. Through this, it performs a three-dimensional analysis of document layouts, text, and context to detect structured personal information patterns. It also precisely detects unstructured personal information hidden in context.
Conventional rule-based and regular-expression-based methods have limitations in handling handwritten text, and they also have limitations in processing information inside tables. They also struggle with areas overlapping seals and stamps, low-quality documents, and rotated documents. By contrast, Lomin Redact effectively handles challenging unstructured documents.
Lomin Redact scores de-identification results by confidence level and selectively presents workers with areas that require review. It also supports the management of de-identification targets through RBAC, or role-based access control. In addition, it supports systematic management of review permissions.
Lenovo Korea plans to use this collaboration as an opportunity to strengthen connections across the domestic specialized AI partner ecosystem. It also plans to continue identifying customized AI adoption cases tailored to enterprise environment optimization.
Yoon Seok-jun, vice president at Lenovo Global Technology Korea, said that as enterprise AI spreads into real-world work environments, data security, governance, and the safety of the deployment environment are essential in addition to computing performance. He added that Lenovo will combine its AI infrastructure capabilities with Lomin's specialized technology to expand safe and practical AI adoption success cases in industries with high regulatory demands, such as finance and the public sector.
Source: IT DAILY · Kwon Young-seok
Original: https://www.itdaily.kr/news/articleView.html?idxno=242305
References
This article was produced with the help of an automated content generation algorithm.
Source: IT DAILY
View originalThis article was summarized and organized by BizCrush based on the original article from IT DAILY. For exact quotations and full details, please refer to the original article.