Technology News on January 2nd, the first working day of 2020, “Top 10 Technology Trends of Dharma Institute 2020” was released. This is the second time that Alibaba’s Dharma Institute has predicted the annual technology trend after 2019.
Looking back at the technological field in 2019, there is still a dark tide surging under the still water. The rise of AI chips, the birth of smart cities, and 5G have given birth to new application scenarios … The technology trends predicted by the Dharma Institute last year have become a reality. The new wave of science and technology has begun in the new decade. Around the fields of AI, chip, cloud computing, blockchain, industrial internet, quantum computing, and other fields, Dharma Institute continues to propose the latest trends and asserts that subversive technological breakthroughs will occur in many areas.
Chip technology has promoted previous waves of technology, but with the slowdown of Moore’s Law and the blowout of high computing power demand scenarios, traditional chips have fallen into the bottleneck of performance growth, and the industry is trying to find ways to crack from all aspects of the chip industry chain. The Dharma Institute believes that breakthroughs in the chip field are likely to be achieved in three areas: architecture, basic materials, and design methods.
In terms of architecture, the von Neumann architecture with separate storage and computing is difficult to meet increasingly complex computing tasks. The industry is exploring an integrated architecture of computing and storage to break through the chip’s computing power and power bottlenecks. In terms of basic materials, silicon is used The representative semiconductor materials are approaching the limit of performance. The sustainable development of the semiconductor industry needs to rely on new materials such as topological insulators and two-dimensional superconducting materials; chip design methods also need to be upgraded, based on chipset modular design Can replace traditional methods, making chip design as fast as building blocks.
Behind the chip technology breakthrough is the “explosion of computing power”, and artificial intelligence is undoubtedly the most important computing power demander and technology puller in the future. At present, the development of perceptual AI technologies such as speech, vision, and natural language processing has reached its limit. However, in terms of cognitive intelligence that leads to “strong artificial intelligence”, AI is still in its infancy. Dharma Institute believes that shortly, AI is expected to acquire autonomous consciousness, reasoning ability, and emotional perception ability, and realize the evolution from cognitive intelligence to cognitive intelligence.
The cognitive evolution of AI has made “swarm intelligence” between machines possible. Dharma Institute predicts that in the future, AI will not only understand “man-machine collaboration”, but also “machine-machine collaboration”. When machines are like humans, they cooperate and compete with each other to complete the target task. It is not difficult to imagine scenarios such as large-scale intelligent traffic light scheduling, warehouse robots collaboratively sporting goods, and autonomous vehicles to sense the global road conditions.
Simultaneously with the transformation of artificial intelligence technology paradigm is the transformation of the IT technology paradigm. Traditional physical machines, networks, and software have stalled. Cloud computing is integrating software, algorithms, and hardware to accelerate digital transformation in all walks of life. The Dharma Institute pointed out that regardless of chip, AI or blockchain, all technological innovations will be centered on the cloud platform. Chips customized for the cloud, AI deeply integrated with the cloud, and blockchain applications on the cloud will continue to emerge. In a nutshell, the cloud will be the center of all IT technology innovation.
The tension between scientific research and application is the eternal driving force for scientific and technological progress. The scientific and technological forecast of Dharma Institute is both forward-looking and fully considered for implementation. Last year, Dharma Institute proposed that the commercialization of blockchain will be accelerated, and this assertion has been verified by reality. In 2019, blockchain technology has become a national strategy and has gradually landed in areas such as digital finance, digital government, and intelligent manufacturing. The Dharma Institute believes that the threshold for enterprises to apply blockchain technology will be further reduced in 2020. Hardware chips such as end-to-end, cloud, and chain-solidified core algorithms designed for blockchain will also emerge. Blockchain applications will enter the masses.
Trend one: the evolution of artificial intelligence from cognitive intelligence to cognitive intelligence
Artificial intelligence has reached or surpassed human standards in the areas of perceptual intelligence such as “listening, speaking, and seeing”, but it is still in its infancy in the field of cognitive intelligence that requires external knowledge, logical reasoning, or domain migration. Cognitive intelligence will draw inspiration from cognitive psychology, brain science, and human social history, and combine technologies such as cross-domain knowledge atlas, causal reasoning, and continuous learning to establish an effective mechanism for stable acquisition and expression of knowledge, so that knowledge can be used by machines Understand and apply, to achieve a key breakthrough from cognitive intelligence to cognitive intelligence.
Trend 2: Integration of computing and storage breaks through the bottleneck of AI computing power
The storage and calculation separation of the Von Neumann architecture is no longer suitable for data-driven artificial intelligence applications. The bottleneck of computing power and the bottleneck of power consumption caused by frequent data transfer have become the limiting factors for the exploration of more advanced algorithms. The in-memory computing architecture similar to the neural structure of the brain integrates the data storage unit and the computing unit into one, which can significantly reduce data handling and greatly improve computing parallelism and energy efficiency. The innovation of computing and storage integration in hardware architecture will break through the bottleneck of AI computing power.
Trend III: Hyper-convergence of the Industrial Internet
[Trend summary] The rapid development of 5G, IoT devices, cloud computing, and edge computing will promote the hyper-convergence of the Industrial Internet and realize the intelligent integration of industrial control systems, communication systems, and information systems. Manufacturing enterprises will implement equipment automation, transportation automation, and production scheduling automation to achieve flexible manufacturing. At the same time, the upstream and downstream manufacturing lines of the factory can adjust and collaborate in real-time. This will greatly improve the production efficiency of the factory and the profitability of the enterprise. For industrial industries with output values of tens of trillions and even tens of trillions, increasing efficiency by 5% -10% will generate trillions of yuan in value.
Trend four: large-scale collaboration between machines becomes possible
Traditional single intelligence cannot meet the real-time perception and decision-making of large-scale intelligent devices. The development of IoT collaborative sensing technology and 5G communication technology will realize the collaboration between multiple agents-machines cooperate and compete with each other to complete the target task. Group intelligence brought by the multi-agent collaboration will further amplify the value of intelligent systems: large-scale intelligent traffic light scheduling will achieve dynamic real-time adjustment, warehouse robots collaborate to complete efficient sorting of goods, and autonomous vehicles can perceive global road conditions. UAV collaboration will efficiently reach the last mile for delivery.
Trend five: modularization reduces the threshold for chip design
Traditional chip design patterns cannot efficiently respond to fast iterative, customized and fragmented chip needs. The open instruction set represented by RISC-V and its corresponding open-source SoC chip design, high-level abstract hardware description language, and IP-based templated chip design methods has promoted the rapid development of chip agile design methods and the open-source chip ecosystem. Also, the modular design method based on chipsets uses advanced packaging methods to package different function “chip modules” together, which can skip tapeouts and quickly customize a chip that meets application requirements, further accelerating chip delivery. .
Trend 6. Large-scale production-grade blockchain applications will enter the masses
Blockchain BaaS (Blockchain as a Service) service will further reduce the threshold for enterprises to apply blockchain technology. Hardware chips such as end-, cloud, and chain-hardened core algorithms designed specifically for blockchain will also be used. It came into being, realizing the anchoring of assets in the physical world and assets on the chain, further expanding the boundaries of the value Internet, and achieving Wanchain interconnection. In the future, a large number of innovative blockchain application scenarios and multi-dimensional collaborations across industries and ecology will emerge. Large-scale production-level blockchain applications with a daily life of more than 10 million will enter the public.
Trend VII: Quantum Computing Enters the Challenge
in 2019, the struggle for “quantum hegemony” has made quantum computing once again the focus of world technology. The achievements of superconducting quantum computing chips have strengthened the industry’s optimistic expectations for the superconducting route and the pace of large-scale quantum computing implementation. In 2020, the field of quantum computing will experience a period of a further increase in investment, intensified competition, accelerated industrialization, and a richer ecosystem. As two of the most critical technological milestones, fault-tolerant quantum computing and demonstrating practical quantum advantages will be the turning point in the practicality of quantum computing. In the next few years, it will be a very difficult task to truly achieve any of them, and quantum computing will enter a period of technical difficulties.
Trend VIII, new materials promote semiconductor device innovation
Under the dual pressure of slowing down Moore’s Law and the explosion of computing power and storage demand, it is difficult for classic transistors based on silicon to maintain the sustainable development of the semiconductor industry. Major semiconductor manufacturers do not tend to chips below 3 nanometers. clear answer. The new materials will implement new logic, storage, and interconnection concepts and devices through new physical mechanisms, driving innovation in the semiconductor industry. For example, topological insulators, two-dimensional superconducting materials, etc. that can achieve lossless electron and spin transport can become the basis for new high-performance logic and interconnect devices; new magnetic materials and new resistive switching materials can bring high-performance magnetics Memory such as SOT-MRAM and resistive memory.
Trend IX: AI technology to protect data privacy will accelerate the landing
Compliance costs from data circulation are getting higher and higher. The use of AI technology to protect data privacy is becoming a new technology hotspot, which can ensure the security and privacy of all parties’ data while cooperating with users to implement specific calculations, solve the problem of data islands and low levels of data sharing, and realize the value of data.
Trend ten: cloud becomes the center of IT technology innovation
With the in-depth development of cloud technology, the cloud has far exceeded the scope of IT infrastructure and gradually evolved into the center of all IT technology innovation. The cloud has penetrated the entire IT technology link of new chips, new databases, self-driven adaptive networks, big data, AI, Internet of Things, blockchain, and quantum computing. At the same time, serverless computing, cloud-native software architecture, software With new technology models such as hard integrated design and intelligent automation operation and maintenance, the cloud is redefining everything in IT. The broad sense of the cloud is continuously turning new IT technologies into accessible services and becoming the infrastructure of the entire digital economy.