[Courtesy of Samsung Electronics]
Advances in hyperscale AI have led to an exponential growth in data volumes that need to be processed, but the performance and efficiency limitations of current computing systems pose significant challenges in meeting heavy computational requirements. Samsung said it would combine its semiconductor design and manufacturing expertise with Naver's experience in the development and verification of AI algorithms and AI-driven services.
Hyperscale AI utilizes massive computing techniques to process big data and cloud computing at lightning speed. Naver has said its hyperscale Korean-language AI platform called "Hyper Clova" can analyzing massive datasets that can be used to develop new AI models.
NAVER will refine its hyperscale language model with over 200 billion parameters, while improving its compression algorithms to create a more simplified model that significantly increases computation efficiency. "We believe we can create an entirely new class of solutions that can better tackle the challenges of today’s AI technologies," Naver Clova CIC head Chung Suk-geun said in a statement on December 6.
Samsung will work with Naver to optimize memory technologies in advancing large-scale AI systems. "Through our collaboration with NAVER, we will develop cutting-edge semiconductor solutions to solve the memory bottleneck in large-scale AI systems," said Han Jin-man, a Samsung executive vice president in charge of memory global sales and marketing.
© Aju Business Daily & www.ajunews.com Copyright: All materials on this site may not be reproduced, distributed, transmitted, displayed, published or broadcast without the authorization from the Aju News Corporation.
![[INTERVIEW] Kakao researcher hopes to develop hyperscale AI models with commercial quality and usability](https://image.ajunews.com/content/image/2022/02/17/20220217154147223151.jpg)
