CUBOX taps HPE to elevate AI innovation for visual recognition

Hewlett Packard Enterprise (HPE) has announced that CUBOX, a growing artificial intelligence (AI) facial and image recognition company in Korea, is developing generative AI models to train its existing visual recognition solutions using an HPE Cray XD supercomputer. The system will also augment CUBOX’s existing innovation focused on developing new image processing technologies.

CUBOX supplies automatic immigration checkpoints to Incheon International Airport in South Korea and facial recognition terminals to four major government buildings, playing an important role in the domestic facial recognition security industry. CUBOX aims to optimize the training of generative AI models by leveraging the power of the HPE Cray XD supercomputer.

Hewlett Packard Enterprise

Transitioning from a business-to-government (B2G) focus on developing facial recognition-based security systems, CUBOX has shifted its attention to business-to-business (B2B) endeavors, particularly in providing AI solutions to the financial sector. Furthermore, the company has embarked on investing in a business-to-consumer (B2C) AI service model, recognizing the pivotal role of generative AI in constructing various B2C service offerings.

“We are driving the evolution of visual recognition technology and shaping a future where AI-powered solutions redefine human interaction to advance security,” stated Un Sung Nam, CEO, CUBOX. “We are building generative AI services based on the new infrastructure to expand our business and apply image recognition solutions to new businesses, such as healthcare, robot systems and digital-twin.”

CUBOX’s system, which is built using the HPE Cray XD6500 supercomputer, is ranked 156 on the TOP500 list of the world’s most powerful supercomputers with a theoretical peak performance of 8.136 petaFLOPs. The boost of computational performance and advanced capabilities that the HPE Cray XD supercomputer delivers has helped support CUBOX provision a technology roadmap for large-scale AI techniques, such as large language models (LLMs) and large multimodal models (LMMs), which are foundational for processing images, text, voice, and videos.

“Supercomputers provide an ideal platform to efficiently support growing compute and data-intensive workloads, such as AI, modeling, and simulation, which require strong scaling and massive performance,” said Young Chae Kim, managing director, Hewlett Packard Enterprise South Korea. “We are honored to power one of the world’s most powerful supercomputers for CUBOX and fuel its mission in advancing visual recognition solutions to drive transformations for government agencies and organizations.”

ChannelDrive Bureau
ChannelDrive Bureauhttp://www.channeldrive.in
ChannelDrive Bureau covers the latest developments in the space of ICT, technology, solutions and implementations and delivers content focused around solution providers, system integrators, distributors and technology partner community in India. ChannelDrive Bureau is headed by Zia Askari. He can be reached at ziaaskari@channeldrive.in

Recent Articles

Related Stories

Stay on op - Get the daily news in your inbox