This article was published in the "Bulletin of the Chinese Academy of Sciences" 2021, Issue 3, "Information and Observation" section, with the original title "Key Points in the Field of Information Technology from 2017 to 2020".
The evolution of information technology is extremely rapid, with significant technological forms such as big data, the Internet of Things, deep learning, and blockchain emerging one after another in the past 10 years, and more emerging technologies are continuously emerging.
At the same time, information technology is profoundly changing human society, and its role in ensuring human welfare has been fully reflected in the prevention and control of the COVID-19 pandemic; against the backdrop of the global economic downturn and the current impact of the epidemic, technologies such as 5G and artificial intelligence are expected to become infrastructure for more industries in the next few years, and like the steam engine and electricity, they will greatly improve social productivity.
Advertisement
In view of the emerging new technologies, new challenges, and new opportunities in recent years, this article has sorted out 7 key trends in the global field of information technology from 2017 to 2020 that are particularly worth paying attention to, including related strategic layout, significant progress, scientific and technological issues, future directions, etc., in order to provide a reference for related strategic judgment and policy deployment.
1
The mainstream technology of artificial intelligence is rapidly evolving, and multiple parties are promoting the "third wave" of artificial intelligence.The mainstream artificial intelligence (AI) algorithm models on the international stage are rapidly evolving, with their efficiency improving swiftly. For instance, the renowned AlexNet's efficiency is known to double approximately every 16 months, and some other representative algorithms see even faster improvements in efficiency. Concurrently, the scale of these leading algorithm models is growing increasingly large. For example, in the field of natural language processing, the parameter scale of leading models ranges from at least 100 million to as high as 1.6 trillion!
Leading institutions and visionaries have begun to promote the "third wave" of artificial intelligence, with the main goal of overcoming the limitations of the second generation of AI, which relies heavily on a large amount of high-quality training data, has low reliability, lacks interpretability, and is deficient in contextual associative reasoning capabilities.
China's Ministry of Science and Technology has laid out brain-like computing and brain-computer intelligence research in its "Science and Technology Innovation 2030 - Brain Science and Brain-like Research" major project. The Chinese Academy of Sciences has also planned brain cognition and brain-like frontier research in strategic pioneering science and technology projects (Class B) and other projects. Tsinghua University has successfully achieved the integration of spiking neural networks with artificial neural networks with its "Tianji Core."
The Defense Advanced Research Projects Agency (DARPA) of the United States has invested a substantial sum of 2 billion US dollars to vigorously promote "next-generation artificial intelligence"; companies such as Google and Open AI in the United States have devoted significant resources to research and development; Intel has scaled up the size of neuromorphic computing systems to 100 million neurons, reaching the level of a small mammal's brain. Carnegie Mellon University in the United States and Google's DeepMind have enhanced the gaming capabilities of AI systems in games such as "Texas Hold'em" and "StarCraft."
(Note: The original text ends abruptly, and the translation stops at the point where the original text ends.)Open-source chips bring new opportunities, DARPA promotes the "Electronic Renaissance" in the post-Moore era
As the demand for chips becomes increasingly diverse with the interconnection of all things and artificial intelligence, open-source chips based on open-source instruction sets such as RISC-V have become a strong new trend. The United States, the European Union, and China have all actively started to layout, and India has even established RISC-V as the de facto national instruction set standard. Currently, the international RISC-V Foundation has more than 500 institutional members, and our country has also established an open instruction ecosystem alliance. The influence of the open-source chip movement at home and abroad is becoming more and more significant, and it is expected to become an important cornerstone of future chip innovation.
The United States, the European Union, South Korea, and others attach great importance to new paths for the development of the semiconductor industry in the post-Moore era, and have successively deployed large-scale strategic R&D plans to promote the next electronic revolution. The "Electronic Renaissance Initiative" led by the US DARPA is the most representative. The plan was launched in 2017 and will invest 1.5 billion US dollars over 5 years, focusing on technical research and development in four key areas: 3D heterogeneous integration, new materials and devices, special function chips, and design and security. DARPA intends to help the US semiconductor industry achieve disruptive innovation and break through existing technical limitations through this plan, thereby ensuring continued control over the global semiconductor discourse.
3
Global quantum competition accelerates, and the heroes compete for the "quantum advantage"
Since 2018, leading countries and regions such as China, the United States, and the European Union have further increased their support for quantum information. At the same time, countries such as the Netherlands, Russia, and Israel have joined in succession, successively launching their own national quantum development strategies, which has significantly accelerated the global quantum competition.Significant advancements in quantum secure communication and quantum computing continue to emerge in both the scientific and industrial sectors. Notably, Google achieved the milestone of "quantum supremacy" in 2019, further stimulating attention and investment in quantum information from various sectors, with a number of large enterprises and startups entering the field. Google's "quantum supremacy" work demonstrated the potential of quantum computing to surpass classical computing, but the "random number sampling" task lacks practicality. In this context, IBM has proposed a more pragmatic measurement standard—"quantum volume." "Quantum volume" takes a more balanced consideration of factors such as the number of quantum bits, the depth of quantum circuits, noise errors, and software compilation efficiency, better reflecting the system's true performance and value, and thus can serve as a reference standard for the development of quantum computing in the future.
4
Brain-computer interfaces are at the "forefront," and the satellite internet competition is fully launched.
In the field of brain-computer interfaces, a comprehensive situation has been formed in recent years with basic research support, military application traction, technological breakthroughs driving, and industrial structure assistance, and the innovation chain and industrial chain have been initially formed, currently at a development "forefront."
The U.S. "Brain Initiative," the European Union's "Human Brain Project," and China's "Brain Science and Brain-like Research" and other large-scale basic research projects have provided strong support for brain-computer interface research to achieve key breakthroughs. DARPA has continuously laid out major brain-computer interface projects, rapidly promoting the progress of related technologies from the perspective of military application needs. Global research institutions are also frequently making technological breakthroughs. The entry of industry giants such as SpaceX and Facebook, along with a number of startups, will accelerate the industrialization process of brain-computer interfaces.In the field of satellite internet, there is a robust market demand, technology is becoming increasingly mature, the policy environment is friendly, and a large number of companies are participating, creating a favorable situation. The global competition in satellite internet has fully started. It should be noted that in our country, satellite internet has been included in the scope of new infrastructure construction, and the future development is promising.
The situation of network security is becoming increasingly severe, and the issue of security prevention is highlighted.
The global network security situation is continuously deteriorating, with major security incidents occurring frequently. For example: the WannaCry ransomware in 2017 caused global panic; the security incidents caused by blockchain in recent years have resulted in losses of up to tens of billions of dollars; artificial intelligence makes hacker attacks more efficient; the complex application scenarios of 5G pose significant threats to Internet of Things (IoT) devices.
Under the new situation, the issue of network security prevention is rapidly highlighted. For example, during the COVID-19 pandemic, the network attacks on China's biosafety laboratories and the World Health Organization surged, and the remote office software ZOOM also exposed major vulnerabilities. In these complex situations, passive security prevention is no longer sufficient, and proactive, intelligent security prevention technology is becoming increasingly important.The role of information technology in supporting significant scientific discoveries is becoming increasingly evident, while the "pitfalls" of information technology continue to emerge.
With the rapid development of information technology itself, its supporting role in major scientific discoveries has become more and more apparent. For example, in November 2020, Google's DeepMind artificial intelligence system, AlphaFold 2, solved a 50-year-old problem that has puzzled biologists - the protein folding problem - at the International Protein Structure Prediction Competition (CASP). It accurately predicted the 3D structure of proteins based on the amino acid sequence of proteins, with an accuracy comparable to experimental techniques such as cryo-electron microscopy, nuclear magnetic resonance, or X-ray crystallography.
At the same time, information technology (especially artificial intelligence) is a "double-edged sword." Currently, there are inherent limitations in artificial intelligence technology itself; at the same time, the misuse of artificial intelligence technology, flaws in data objects, and other factors have led to the continuous emergence of "pitfalls" in its application to scientific research.
For example, in 2019, Nature magazine reported on two material research and earthquake research studies completed with the help of artificial intelligence technology, but artificial intelligence experts found serious errors in the use of artificial intelligence technology, making the results unreliable. After the outbreak of the COVID-19 pandemic globally, computer scientists at home and abroad actively proposed artificial intelligence solutions for the new coronavirus. However, many of these solutions lack clinical experience or medical capabilities, or use flawed data, providing no help to patients and doctors.
The information industry has become a focus of competition among powerful countries, and the industrial and innovation ecosystems are being reconstructed.Countries and regions that have mastered the core competitive advantages of the information industry are engaged in fierce competition around the new round of international division of labor. Since 2018, the United States has initiated trade frictions with our country and dozens of other countries; at the same time, the United States has adopted a "decoupling" policy towards our high-tech, especially the information technology field, including technological blockade, industrial supply interruption, market entry ban, talent flow interruption, academic disconnection, and global containment, etc. This has severely damaged international cooperation and exchange in related aspects and has dealt a heavy blow to the global industrial chain. Reconstructing the industrial and innovation ecology in related fields has become one of the urgent key tasks at present.
In 2019, the United States' restrictions on our country were mainly related to hardware, including artificial intelligence and machine learning, microprocessors, robots, quantum computing, brain-computer interfaces, data analysis, advanced computing, positioning and navigation, etc.; in 2020, it expanded to the field of software, including geospatial image software, point cloud analysis software, satellite image analysis software, 3D spatial technical analysis software, etc. In these directions, the original global industrial and innovation ecology is being reconstructed.