Visit the purchase page:
The picture is from NeoWin
For future development planning, Inyup Kang, president of Samsung Electronics Systems LSI business, said: "For the coming era of artificial intelligence, Samsung is committed to leading the industry through our NPU technology. As we use differentiated technologies, build close partnerships with global institutions and actively invest in top talent, we are pleased to be able to upgrade our future AI processing capabilities to a new level.
With the help of a new generation of NPU technology, Samsung will develop hardware for on-board information entertainment system (IVI), Advanced Driver Assistance System (ADAS) and the next generation of data centers for optimizing the processing of large data.
Samsung Electronics launched xinos 9 (9820) last year with an independent neural network processing device. This is the result of ongoing research in the LSI Division of Systems and the Institute of Integrated Technology. This product is the basis of the system semiconductor "System-level Chip (SoC)", which embodies many functions in one chip. It realizes "in-device artificial intelligence" and can be completed by itself in a mobile machine before the need to pass andCloud servicesArtificial Intelligence Operations for Data Interaction.
The advantage of in-device AI is that data does not pass through the cloud, so it is excellent in protecting personal information. Similarly, because it does not pass through the cloud, the answer speed is very fast, and in the absence of a network environment, it can also be driven by low power and low cost. The company expects that new areas of inference and learning will emerge if it combines the fifth generation (5G) mobile communication technology with cloud and peripheral AI machines.
Zhang Dexian, Director of SoC Development Room, LSI Division of Systems, said: "In the future, almost all IT products and services will have built-in AI functions. Based on this trend, the application of neural network processing devices is not only in intelligence.Mobile phoneOn the other hand, it will expand to automatic driving, data center, cloud and so on. It is predicted that the annual growth rate will be 52%.
Samsung Electronics plans to expand the installation of neural network processing devices in the whole IT field, such as electrical equipment, data centers and the Internet of Things (IoT). Firstly, starting from the flagship SoC products for mobile communications, we concentrate on the development of automobile SoC products, such as information entertainment system and advanced driving assistance system (ADAS). Their goal is to further develop the processing technology of neural networks, and to realize the processing technology of information processing and recognition at the level of human brain.