The preface
In the era of steady progress, we use technology to explore the “unknown” of the world; In times of uncertainty, we anchor the “certainty” of the world with technology.
Right now, our mission leans toward the latter. In 2022, as the COVID-19 pandemic continues to impact and the global economy faces many challenges, science and technology have become an important force driving innovation and development. Artificial intelligence, which combines technological development and industrial value creation, features the rapid evolution of core technologies, enhanced cross-field connectivity, and an increasingly solid industrial base, is just such a scientific and technological force.
First of all, AI core technologies continue to make breakthroughs and evolve, and integration and innovation are becoming more and more significant. For example, knowledge enhancement, cross-modal, cross-language and other technical paths activate the imagination space of the large model, and even ignite the hope of general artificial intelligence.
Second, in interdisciplinary and interdisciplinary research, AI constitutes a universal variable in scientific research and technological development. In quantum, biology, chemistry and other fields, “+AI” has become one of the most exciting research directions.
Third, in terms of value creation, AI is driving the development of autonomous driving, robotics, aerospace, life and health, and playing an important role in achieving the goals of “dual carbon” and universal access to science and technology.
Today, Baidu Research Institute shares with you the forecast of the technology trend in 2022, hoping to light up the road of innovation with AI as the lamp in the uncertain era. Take AI as an oar and start the boat of development.
— Wang Haifeng, CTO of Baidu and President of Baidu Research Institute
1. The super-scale pre-training model presents the trend of knowledge enhancement, cross-modal unified modeling and co-evolution of multiple learning modes, and gradually becomes practical
Based on massive data, the large model conducts self-supervised learning and uses unified models and paradigms to solve various AI tasks, breaking the dependence of traditional technologies on large-scale annotated data and significantly improving the effect, universality and generalization of AI models.
Is expected in 2022, large model research and development direction will shift from continue to increase the size of the parameters to the practical application, based on the knowledge enhancement, cross modal unified modeling, prompt, continuous learning, combining model of distillation, sparse technology, such as the effect of the large model, generality, generalization, interpretability and will continue to improve operation efficiency, and application of the threshold, the decrease of In this way, the Internet, intelligent office, intelligent finance and other scenarios are widely implemented. For example, AIGC (AI generated content) can stimulate creativity, improve content diversity and reduce production costs with the help of cross-modal comprehensive technical capabilities of large models, which will achieve large-scale application.
The emergence of new research fields of AI for Science is expected to bring about paradigm changes in scientific research
Machine learning help mathematician found two hypothesis, and multi-scale modeling and high performance computing using machine learning, a combination of quantum random simulation circuit to solve large scale problems, let people see the artificial intelligence applied to scientific research, in processing data, design a new experiment, and create a more efficient calculation model of huge potential.
The emerging AI for Science is expected to promote the deep integration of the two research paradigms of data-driven and theoretical deduction. In the coming years, AI is expected to further integrate with different fields such as mathematics, physics, chemistry, materials and engineering, and play a bigger role in the progress of basic science.
3. Ai-based biocomputing will continue to develop rapidly, and new breakthroughs will be made in the collaborative innovation of basic research and application scenarios
In the context that human society is still fighting against novel Coronavirus, the life and health industry’s demand for technological innovation is more urgent. AI has enabled gene editing to find targets more accurately and quickly, and AI has helped make significant breakthroughs in protein structure prediction. The success of THE NOVEL coronavirus mRNA vaccine technology has led to the explosion of drug design and vaccine research and development based on RNA, protein and other macromolecules, and international mainstream drug manufacturers are accelerating the implementation of mRNA technology.
In the future, ai-based biocomputing will make breakthroughs in more basic research and application scenarios, such as protein-based drug design, synthesis and screening, anti-cancer drugs based on mRNA technology, monoclonal antibodies and immunotherapy, etc. The deep integration of the two will significantly shorten drug development cycles, reduce r&d costs, and promote precision medicine and personalized diagnosis and treatment.
4. Privacy computing technology has attracted much attention and will become a breakthrough for releasing data value and an infrastructure for building trust
With the improvement of global personal information and data security regulations, it has become a consensus in the industry that security compliance is the premise to promote the effective release of data value.
Privacy computing technology, represented by trusted confidential computing and federated computing, has attracted much attention because of its consideration of data security and data sharing. With the improvement of the performance of privacy computing technology, the mutual promotion of technology and compliance standards, and the cooperation of multiple parties to improve the credibility of technology, relevant typical applications will appear in the scenarios of biological computing, financial analysis and data transaction.
In the long run, private computing may push the flow of data and computing based on cryptography as the default option, becoming an infrastructure for building trust.
5. The integration of quantum software and hardware has become the mainstream trend, and the realistic demand is to accelerate the integration and innovation of quantum computing with various industries
It is expected that in 2022, the design, preparation and measurement and control technology of quantum chips will continue to develop, the number of quantum bits will realize the scale growth, and seek breakthroughs along the two ideas of noise reduction or adaptation. With the development of quantum software and services across platforms, users will get richer choices of quantum backend on cloud native quantum computing platform, and the quantum platform bearing quantum soft and hard integration scheme will gradually show its application value.
With the deep integration and innovation of quantum computing and intelligent manufacturing, artificial intelligence, chemical medicine, financial technology and other fields, a number of practical application solutions with significant quantum advantages will emerge one after another. Government agencies, research institutes and the industry will also work more closely to build high-quality quantum equipment and train quantum scientific and technological talents, and initially open up the quantum computing industry chain.
6. Autonomous driving technology has entered a new stage of unmanned landing, and multiple “car robots” continue to emerge, connecting technologies and scenes
In 2022, under the dual impetus of policies, regulations and technological progress, autonomous driving will advance rapidly in the unmanned, diversified “car robot” represented by the rapid development of the car form.
Through the application of passenger cars, buses, trunk logistics, warehousing and distribution, special operations in mines and ports, retail, sanitation and other rich scenarios, multiple “car robots” will provide more extensive services for users, create value for customers, and gradually realize steady business income, promote the development of science and technology and social progress.
7. Integration and innovation of AI technology and aerospace science and technology will push deep space exploration to a new stage of intelligence
Deep space exploration carries mankind’s curiosity and imagination about the universe and itself. To realize the moon and planet stay, to carry out scientific exploration and resource development and utilization as the main program, to carry out deep space exploration in remote and unknown environment, the need for the autonomy of probes is increasingly strong.
In the field of construction machinery automation, the actual engineering scene landing of 24-hour continuous unmanned excavation operation has been realized. Related AI algorithms such as autonomous environment perception and motion planning will also enable the detector to have the functions of autonomous obstacle avoidance and decision-making, flexible and autonomous operation of mechanical arm in the future. In addition, AI technology is also expected to play an important supporting role in spacecraft fault detection and repair, the construction of digital twin simulation laboratory, and deep space big data detection and analysis.
8. “Social distance” accelerates human-machine symbiosis and supports the rapid integration of virtual and real technology and intelligent interaction into production and life
The COVID-19 pandemic has created a “social distance” for people’s communication, and the development of digital technology allows us to shorten this distance, accelerating the symbiosis between humans and digital humans and robots. The future world of virtual reality and intelligent interaction is no longer far away from us.
This change is supported by the continuous progress of AI technologies such as vision, voice, natural language processing and XR in cross-modal understanding, generation and continuous learning, as well as the cross-technology support system formed by integrating hardware, network, computing, ecosystem platform and content.
With the acceleration of the integration and innovation of relevant technologies and the maturity of the cross-technology support system, more virtual and real combination and intelligent interactive products will emerge for industry and consumption scenes, thus promoting the deep integration of digital economy and real economy and enriching people’s production and life experience.
9. Green and low-carbon will be more integrated into the AI blueprint to help achieve peak carbon neutrality
Data centers and large-scale AI computing have realized important economic and social values, but their energy consumption and impact on the environment cannot be ignored. There is an urgent need to develop more environmentally friendly “green AI” technologies to reduce the energy consumption of model training and use.
In the coming years, “green AI” related technologies will continue to flourish, building systems around energy-efficient architecture design, training and reasoning strategies, data utilization, etc., and forming evaluation standards that take into account both performance and energy consumption. AI chips with higher computing power and lower energy consumption will continue to emerge; Leading AI enterprises build intensive large computing power and large model to improve downstream performance and reduce overall energy cost; Policies will also encourage the construction of green and low-carbon data centers and the introduction of AI technology to improve the energy efficiency of infrastructure.
10. AI is more inclusive and beneficial to all. The value creation orientation brings more attention to the needs of small and medium-sized enterprises and vulnerable groups
Pratt & Whitney AI is not only about AI practitioners, but also about the broader beneficiaries of AI technology.
Open source platforms with deep learning frameworks as the core have greatly lowered the threshold for AI technology development. Public data sets, large model bases and regional intelligence and computing centers will be further developed to help smes reduce costs and increase efficiency and stimulate innovation. A nationwide AI training system will also be gradually established to promote the re-employment of personnel in traditional industries and AI popular science education.
AI should also benefit the welfare of all social groups, along with the policy guidance and the sustainable development of the concept of ESG, attention will turn to value creation, AI service demand for old people, children and other vulnerable groups will strengthen the attention, develop the corresponding pratt &whitney AI services and products, let everybody can enjoy the convenience of digital technology.
