Top ten technology trends in 2019, including: real-time simulation of the city became possible, the smart city was born; voice AI passed the Turing test in specific areas; AI-specific chips will challenge the absolute rule of the GPU Status; hyperscale graph neural network system will give machine common sense; computing architecture will be reconstructed; 5G network will spawn new application scenarios; digital identity will become the second ID card; automatic driving enters a cool development period; blockchain returns to rationality , commercial application acceleration; data security protection technology is accelerating.

The following are details of the 2019 Top Ten Technology Trends:

Trend 1: Real-time simulation of cities is possible, the birth of smart cities


The city’s public infrastructure’s perceptual data and urban real-time pulsating data streams will be aggregated on the big computing platform. The development of computing power and algorithms will promote the real-time integration of unstructured information such as video and other structured information. Urban real-time simulation becomes possible. Intelligence will be upgraded to global intelligence. In the future, more power will be developed for the research and development of urban brain technology and applications. On the physical city, a smart city with full-time perception, full-element linkage, and full-cycle iteration will be born, which will greatly promote the optimization of urban governance. Ascension, it is expected that in the new year, more and more cities in China will have brains.

Trend 2: Voice AI passes the Turing test in specific areas

With the standardization and low cost of the end-cloud integrated voice interaction module, there will be more and more talking public facilities. In each future, there will be at least one contact for voice interaction. With the advancement of intelligent voice technology, real-time voice generation on mobile devices and live voice may be indistinguishable, even through Turing tests in some specific conversations. Rules and even laws in this area will be gradually established to guide the industry towards standardization.

Trend 3: AI-specific chips will challenge the absolute dominance of GPUs

In the current AI training scenario of the data center, data migration between computing and storage has become a bottleneck, and a new generation of AI chip architecture based on 3D stacked storage technology has become a trend. The demand for data bandwidth in AI chips will further promote the widespread application of 3D stacked memory chips in AI training chips. Brain-like computing chips will further promote their development in finding more suitable applications. In the training scenario of the data center, the AI-specific chip will challenge the absolute dominance of the GPU. The AI chip architecture that truly embodies Domain Specific will still be more reflected in many edge scenarios.

Trend 4: Hyperscale graph neural network system will give machine common sense

Pure deep learning has matured, and the combination of end-to-end learning and inductive reasoning combined with deep learning graph neural network is expected to solve a series of problems such as relational reasoning and interpretability that deep learning cannot handle. A powerful graph neural network will be similar to the brain of a person formed by a network of nodes such as neurons. The machine is expected to become an AI with common sense and understanding and cognitive ability.

Trend 5: The computing architecture will be refactored

Whether in a data center or edge computing scenario, the computing system will be refactored. Future computing, storage, and networking must not only meet the demand of artificial intelligence for high-throughput computing power, but also meet the demand for low-power consumption in the Internet of Things. The heterogeneous computing architecture based on FPGAs, GPUs, ASICs and other computing chips, as well as the emergence of new types of memory devices, has begun the evolution of traditional computing architecture. From the past CPU-centric general-purpose computing to the disruptive change of the Domain-specific architecture brought about by application-driven and technology-driven, it will accelerate artificial intelligence and even quantum computing gold. The arrival of the times.

Trend 6: 5G network spawns new application scenarios

The fifth-generation mobile communication technology will greatly enhance the mobile bandwidth, providing nearly 100 times the peak rate of 4G, and promote the gradual maturity of immersive interactive mode based on 4K/8K ultra-high definition video and AR/VR. The connectivity will be enhanced to tens of billions, bringing a massive integration of machine-like communications and connectivity. The network evolves into a cloud and software. The network can be sliced into multiple independent and parallel virtual sub-networks to provide virtual private networks for different applications. Together with high reliability, low latency, and large capacity network capabilities, the network will enable vehicles. Road synergies, industrial Internet and other fields have gained new technological empowerment.

Trend 7: Digital identity will become the second ID card

Biometrics technology is maturing and entering a large-scale application phase. With the rapid spread of 3D sensors and the integration of multiple biometrics, each device can “see” and “listen” smarter. Biometric and in vivo technologies will also reshape identity and certification, and digital identity will become the second identity card for people. From the unlocking of mobile phones, community access control to restaurant meals, supermarket cashiers, to high-speed rail stations, airport security inspections and hospital visits, the era of relying on the face of the world is accelerating.

Trend 8: Autopilot enters a cool development period

Relying on the “cycling intelligence” approach to revolutionizing the car, the ultimate driverlessness can’t be achieved for a long time, but it doesn’t mean that the automatic driving is completely into the cold winter. The road-to-technical route will speed up the arrival of driverless. In the next 2-3 years, the commercialization of autonomous driving, represented by limited scenes such as logistics and transportation, will usher in new progress. Commercial scenes such as fixed-line bus, unmanned distribution, and park micro-circulation will quickly land.

Trend 9: Blockchain regression rationality, commercial application acceleration

In the digitalization process of various industries, the Internet of Things technology will support the trusted mapping of the world under the chain and the data on the chain. The blockchain technology will promote the reorganization and optimization of trusted data in the circulation path, thereby improving the efficiency of circulation and coordination. . In many scenarios, such as cross-border remittances, supply chain finance, electronic bills and judicial deposits, blockchains will begin to integrate into our daily lives. With the value of “link”, layered architecture and cross-chain interconnection will become the technical basis for blockchain scale. The blockchain field will return to rationality from excessive fanaticism and excessive pessimism, and commercial applications are expected to accelerate.

Trend 10: Data Security Technology Accelerates

Governments will tend to introduce more stringent data security policies and regulations, and companies will invest more in personal data privacy protection. In the next few years, hacker and black production attacks will not stop, but data security protection technology will be launched in addition. Cross-system data tracking and traceability related technologies, such as watermarking technology, data asset protection technology, and advanced anti-reptile technology for strong confrontation, will be more widely used.