What is the new development trend of artificial in

What are the new development trends of artificial intelligence in the future

What are the new development trends of artificial intelligence in the future

16:01 source://

original title: what are the new development trends of artificial intelligence in the future

AI chips include three major markets: Data Center (cloud), communication terminal products (), and specific application products (self driving, head mounted AR/VR, UAV, robot...). At present, there are eight new trends in AI development in the future

trend 1: AI has great potential in vertical fields of various industries

artificial intelligence market has great potential in vertical fields of various industries such as retail, transportation and automation, manufacturing and agriculture. The main factor driving the market is the increasing number of applications of artificial intelligence technology in various end-user vertical fields, especially to improve the service to end consumers

of course, the artificial intelligence market will also be popularized by the improvement of IT infrastructure, intelligence and intelligent wearable devices. Among them, natural language processing (NLP) application market accounts for a large part of AI market. With the continuous improvement of natural language processing technology to drive the growth of consumer services, there are also fields such as automotive information communication and entertainment system, AI robot and AI supporting intelligence

trend 2: AI is introduced into the health care industry to maintain rapid growth

due to the extensive use of big data and artificial intelligence in the health care industry, it can accurately improve disease diagnosis, manpower imbalance between medical personnel and patients, reduce medical costs and promote cross industry cooperation

in addition, AI is also widely used in clinical trials, large-scale medical plans, medical consultation, publicity and promotion, and sales development. Artificial intelligence has been introduced into medicine. With the rapid improvement of people's favor for LED lighting system in recent years, the medical and health care industry has maintained a high growth from 2016 to 2022. It is expected to reach $7988.8 million in 2022 from $667.1 million in 2016, with an average annual compound growth rate of 52.68%

trend 3: AI replaces the screen as a new UI/UX interface

since the PC era, user interfaces have interacted through the screen or keyboard. With the smart speaker, virtual/augmented reality (VR/AR) and autonomous vehicle systems entering the human living environment one after another, people can easily communicate with the computing system without a screen

this means that artificial intelligence makes technology more intuitive and easier to manipulate through natural language processing and machine learning. In the future, it will replace the screen in user interface and user experience

in addition to playing an important role in the enterprise back-end, AI can also play a more complex role in the technical interface. For example, automatic driving using visual graphics has the advantages of high efficiency, energy saving and high cost performance. The vehicle can realize real-time translation through artificial neural network. In other words, artificial intelligence makes the interface simpler and more intelligent, so it sets a high-standard mode of interaction in the future

trend 4: non exterior wall insulation materials have been pushed to the cusp of the storm. The chip must have built-in AI computing core

at this stage, the mainstream arm architecture processor is not fast enough. It is still insufficient to carry out a large number of image operations, so the chip in the future will have built-in AI computing core. For example, after Apple introduced 3D sensing technology into the iPhone, Android smart camp will follow up the introduction of 3D sensing related applications next year

trend 5: the key to AI chips lies in the successful integration of software and hardware

the core of AI chips is semiconductors and algorithms. AI hardware mainly requires faster instruction cycle and low power consumption, including GPU, DSP, ASIC, FPGA and neuron chip, and must be combined with deep learning algorithm. The key to successful combination lies in advanced packaging technology

generally speaking, GPU is faster than FPGA, and FPGA is better than GPU in power efficiency, so the choice of AI hardware depends on the needs of product suppliers

for example, Apple's faceid face recognition is a 3D depth sensing chip with neural engine operation function, which integrates up to eight components for analysis, including infrared lens, floodlight sensing component, distance sensor, ambient light sensor, front-end camera, dot matrix projector, speaker and microphone. Apple emphasizes that users' biometric data, including fingerprint or face recognition, are stored inside the iPhone in encrypted form, so it is not easy to steal

trend 6: AI autonomous learning is the ultimate goal

AI "brain" becoming smart is carried out in stages, from machine learning to deep learning, and then to autonomous learning. At present, it is still in the stage of machine learning and deep learning. To achieve autonomous learning, four key problems need to be solved

first, build an AI platform for Autonomous machines; We should also provide a virtual environment that allows Autonomous machines to learn independently. It must comply with the laws of physics. The impact, pressure and effect should be the same as those in the real world; Then put AI's "brain" into the framework of autonomous machine; Finally, a virtual world portal (VR) is established

at present, NVIDIA has launched its own machine processor Xavier, which is preparing for the commercial and popularization of independent machines

trend 7: the most perfect architecture is the combination of CPU and GPU (or other processors).

in the future, there will be super performance processors required in many special fields, but CPU is common to all kinds of devices and can be applied to any scenario. Therefore, the most perfect architecture is the combination of CPU and GPU (or other processors). For example, NVIDIA combines the common computing model with the special computing model

trend 8: AR becomes the eye of AI. The two are complementary and indispensable

Future AI needs AR, and future ar also needs AI. AR can be compared to the eye of AI. The virtual world created for robot learning is virtual reality itself. In addition, if people want to enter the virtual environment to train robots, more other technologies are needed

looking forward to the future, with the maturity of AI, IOT, VR/AR, 5g and other technologies, it will drive the 30-year prosperity of a new wave of semiconductor industry, including four chips: memory, central processing unit, communication and sensor. The demand for application chips of various new products is increasing. With China's huge market advantage in semiconductors, it can definitely play a key role in the world

this article was originally published by @ Xie Jun. everyone is a product manager. Without permission, it is forbidden to reprint

Title drawings from pexels, based on cc0 protocol