As Richard Kingston, vice president of market information at CEVA, says, human society is entering the fourth industrial revolution, the same as the previous three industrial revolutions (mechanization, mass production and electrification, computers and automation), and the fourth industrial revolution will once again be Change people's lives.

Human beings landing on the earth used to be the science fiction of the ancients. It has already been realized. Today we are faced with the scene of autonomous driving and human-machine conversation. We may wish to boldly envision the future.

The transformation of knowledge and information forms is the most important part of how the fourth industrial revolution will change the future. In the past, knowledge was written on paper, kept in universities and libraries, and processed and used by people. Today, knowledge is stored in modern data centers, and various machines and equipment analyze and use them. Kingston said: "This is a huge shift in the digitalization process, and the world has changed."

From financial services and healthcare to retail and home, there are huge amounts of information and data based on cloud computing in many industries and applications. The question is, who should handle this information – is the cloud still a terminal with artificial intelligence ? In fact, these two methods will coexist. Kingston added: "The future terminals will automate data processing and decision making. More importantly, they will continue to learn and feed the results back into the cloud's knowledge base, and they can also launch next-generation networks for specific end markets. Recently I have seen Some data pointed out that by 2022, about 50% of AI devices will have built-in machine learning capabilities - if the data is correct, this would be a considerable amount."

Front-end AI - the peripheral nerve of artificial intelligence

Having artificial intelligence at the far end helps solve the following problems:

Latency: Especially important in a security-first scenario

Privacy: Devices should not send private data to the cloud

Security: Cloud data is more vulnerable to hackers

Network coverage: no network connection to the cloud

Cost problem

Kingston continued: "If you rely on connecting to the cloud to perform a lot of AI processing, this will cause a series of problems. In the automotive industry, for example, processing latency is a big problem, and you don't want to have private data. Stored in the cloud." So cloud.

Front-end AI has a large number of application scenarios, such as:

Intelligent surveillance, face detection, voice biometrics, voice detection, motion sensing, connectivity, automotive electronics, vision sensors, communications, radar and lidar, GPS, connectivity – Bluetooth, WiFi, Cellular networks, data fusion and data analysis

Since the front-end AI has been added to smartphones, neural network processors and hardware accelerators have gradually become mainstream technologies, and AI remote processing has taken a big step widely adopted by the industry. Both Apple and Huawei have introduced a dedicated neural network processor for face recognition in their end products. This solution, all completed by the terminal, has the advantages of security, privacy and directness.

Qualcomm and NVIDIA also released neural network processors for smartphones and other mobile devices. CEVA predicts that each camera-equipped device will have built-in visual processing and neural network processors for artificial intelligence in a few years. By 2020, one-third of smartphones will support artificial intelligence.

In addition, data processing performance and power consumption efficiency are expected to increase tenfold in the same period, which is crucial, because the upgrade demand for data processing performance of smart phones is much higher than the development speed of battery capacity. Kingston went on to say: "The development of battery capacity and (data) processing capacity upgrades are huge, and this will only get worse. When you need to add a neural network processor to these devices to run AI functions in real time, You will consume battery power faster than ever before. If the battery technology continues to grow at the same speed, your device will not be able to stand by for a long time. Therefore, it is necessary to solve the problem both in terms of processing power and battery technology. ".

Front-end AI involves four main components—communication, sensor processing, data fusion and data analysis, and application/implementation. “CEVA provides solutions for the first three parts.”

“Looking at the big picture, we focus on the three challenges facing the front-end AI market – power consumption, price and soaring performance requirements.” CEVA achieves this goal through technology licensing to a wide range of devices from the terminal to the base station. Includes sound equipment, connected devices, and wireless network devices.

Regardless, the current focus is on vision systems, including computer vision DSPs, neural networks, accelerators that can handle different application scenarios, from cell phones to cars, and neural network frameworks.

The number of global cameras is expected to increase by 216% from 2016 to 2022, when there will be approximately 44 billion devices with camera-enabled. This new vertical growth trend is due to the fact that more and more different devices have begun to integrate cameras.

CEVA offers a total solution for CEVA-XM vision DSP for computer vision and artificial intelligence workloads, as well as a library of image processing and video processing software, CEVA deep neural networks and hardware accelerators to provide a scalable visual and deep learning solution. Program. The CEVA-XM Computer Vision and Neural Network Ecosystem is an open platform that has acquired a large number of licenses covering end markets including drones, surveillance, automotive electronics, smart homes and robotics.

CEVA also recently announced the development of a smart 3D camera with LG, which will use a multi-core CEVA-XM4 vision DSP. At the same time, LG effectively reduces costs by deploying self-developed algorithms on existing commercial chips. CEVA has also entered into a partnership with startup Brodmann 17, which develops deep learning software on embedded devices that uses CEVA-XM processors at processing speeds of up to 100 frames per second. This is 170% faster than NVIDIA's Jetson X2 platform.

Kingston concluded: "We will see more forms of artificial intelligence devices on the market, including smartphones, drones, ADAS and surveillance equipment. CEVA is working on product development for the next generation of computer vision and neural networks, especially Unlimited neural network technology and how to build the most effective network training system in the data center.

“We hope to continue investing in the artificial intelligence ecosystem based on the current products, and we hope to purchase some additional technologies to enrich our products, which will help customers enter the market faster. CEVA is no longer a DSP only. Nuclear communications companies, now we offer a wider range of technologies for different fields, let us wait and see for the next few years!"

PE Terminal Block

Terminal Block Din Rail,Spring Grounding Terminal Block,Pe Grounding Terminal Block,Plastic Terminal Blocks

Wonke Electric CO.,Ltd. , https://www.wkdq-electric.com