Samsung and Naver will develop a chip optimized for large-scale AI

South Korean Samsung, and the South Korean technology giant Naver, announced on December 6, 2022 their collaboration. Together, the two firms wish to develop semiconductorsspecifically chips, suitable for the most successful artificial intelligence (AI) models and capable of being exploited at large scale.

A dedicated chip for creating and supporting large-scale AI models

If artificial intelligence may be the solution problems present within certain business departments, its adoption is often complicated. The goal of large-scale AI is to enable growing businesses and organizations digital transformation to have access to powerful tools while facilitating their adoption at all levels.

Large-scale AI models must be generally complete, able to process a very large amount of data, without being too complex to understand, while being close to the systems already existing within a company so as not to confuse a business direction. . ” Developing such solutions requires extensive convergence of semiconductor and AI disciplines “, also specifies Samsung in his press release.

In order to help developers develop such solutions, while offering something powerful, Samsung and Naver decided to join forces to design a chip dedicated to these needs. The two firms will use their respective skills in the AI ​​and semiconductor sectors. If companies like Nvidia are used to designing high-performance electronic components for uses related to AI, this is not the case for Samsung, which has logically partnered with Naver to help it in its task.

Samsung and Naver put their respective skills to good use

In order to carry out this project, the engineers who will work on the design of this chip will be able to rely on the various storage and memory technologies developed by Samsung in recent years. This will support the processing of a large amount of data at high speed. Of course, the South Korean giant also knows about electronic components. It has also started the production of its first chips engraved in 3nm in June 2022.

The actors of this project will also rely on NAVER’s artificial intelligence skills. The firm is known for having developed HyperCLOVA, a large-scale language model composed of 200 billion parameters, 25 billion more than OpenAI’s GPT-3. The South Korean company should seek to modify HyperCLOVA in order to develop an even more simplified model of it, which would increase the efficiency of the calculations of the chip soon created.

This is not the first time that the two entities have collaborated on the implementation of a major project. Between 2020 and 2021, Samsung and Naver had worked together to develop the first network 5G from South Korea, within Naver’s own headquarters. This had allowed the firm to operate autonomous mobile robots on a daily basis with all data based in the cloud and at the edge.

We wish to thank the writer of this write-up for this amazing web content

Samsung and Naver will develop a chip optimized for large-scale AI


Discover our social media profiles , as well as the other related pageshttps://www.ai-magazine.com/related-pages/