Guide
On September 24, 2023, the 2022/2023 China Urban Planning Annual Conference opened in Wuhan. At the second session of the special session "New Technology Empowerment Planning", Professor Zhen Feng gave a report entitled "AI-driven Urban Research and Planning Thinking", which mainly shared from three aspects: the transformation of scientific research paradigm from data-driven to AI-driven, the transformation and development of urban planning and innovation needs, and the prospect of AI-driven urban research and planning.
Number of characters: 3015 characters
Reading time: 10 minutes
Zhen Feng
Member of the Special Committee on the Application of New Technologies in Urban Planning,
Vice Dean and Professor of School of Architecture and Urban Planning, Nanjing University
01
Paradigm shift in research: from data-driven to AI-driven
1.1 Human society has entered the era of intelligent society with artificial intelligence as the main driving force
Digital China and smart cities are important engines for promoting Chinese-style modernization in the digital era, using a new generation of information technology to empower the modernization of national governance and high-quality development, and become an important driving force for promoting sustainable development and new urbanization in the context of ecological civilization. The 2015 Central Urban Work Conference pointed out that the level of planning should be improved, and the scientific and authoritative nature of planning should be enhanced, and the 2018 meeting of the Central Committee for Comprehensively Deepening Reform proposed that modern information technologies such as big data and cloud computing should be comprehensively used to innovate the means of planning and preparation.
1.2 The scientific and technological revolution empowers scientific research
Artificial intelligence (AI) is a disruptive technology (Batty, 2018, Yigitcanlar et al., 2020), which has widely penetrated into the fields of natural sciences and humanities and social sciences, injecting new elements and new momentum into scientific research, changing the scientific research model, improving the efficiency of scientific research, and giving birth to a new scientific research paradigm.
GPT technology may disrupt the order of traditional knowledge production in the humanities and social sciences. Traditional knowledge production has formed a relatively complete system of "scholar-centered". And the general large model represented by GPT technology may break our traditional knowledge production order. (Gao Qiqi, 2023)
AI technology not only greatly improves the efficiency and accuracy of common tools in scientific research activities, but also helps to establish an effective system for scientific research driven by industrial needs, promotes the transformation of scientific research mode from scholars (teams) to "platform scientific research" mode, and empowers practical scenarios of urban governance, industrial development and life applications.
Interdisciplinarity becomes more possible. Previous interdisciplinary research was extremely difficult because different disciplines created corresponding knowledge barriers. Both the research methods of technical theories and the discourse system are completely different.
1.3 Paradigm shift from data-driven to AI-driven research
Jim Gary believes that scientific research has gone through four paradigms: empirical paradigm, theoretical paradigm, computational paradigm, and data-driven paradigm. At present, many scientists believe that scientific research is ushering in a new paradigm, that is, the fifth paradigm, that is, taking virtual and real interaction and parallel-driven AI technology as the core, building the foundation of intelligent networking and blockchain, and considering the integration of human value and knowledge (Wang Feiyue, Miao Qinghai, 2023).
For example, in early 2020, the U.S. Department of Energy released the "AI for Science Report" to promote the application of artificial intelligence in science, covering fields from high-energy physics and materials science to computing technology, and in March 2023, the Ministry of Science and Technology, together with the Natural Science Foundation of China, recently launched the special deployment of "AI for Science", which closely combines key issues in basic disciplines such as mathematics, physics, chemistry, and astronomy, focusing on drug research and development, genetic research, biological breeding, and The scientific research needs in key areas such as the research and development of new materials have been expanded, and the cutting-edge scientific and technological research and development system of "artificial intelligence-driven scientific research" has been laid out.
At present, the application of machine learning technology in social science research can be divided into three categories:
01
Data Generating Process: Machine learning can help academics obtain data that was previously difficult or impossible to obtain;
2
Prediction: Machine learning can more effectively explore the correlation between variables to make more accurate predictions.
3
Causal Inference: Causal recognition is at the core of empirical research in the social sciences, especially in economics, and machine learning also has certain advantages in this regard.
Data:
Active learning, multi-task learning, new data production methods, solve problems such as insufficient data.
Algorithm:
Used to solve a specific problem or perform a specific task. Algorithms can describe the logical flow in a computer program or they can be used to describe the steps a human takes while solving a problem.
machine
Learn
Model:
An algorithm that uses data for training that predicts or decides on an output given an input. Common types of machine learning models include: supervised learning models, unsupervised learning models, reinforcement learning models, etc., and different models have their own in-depth theory and practice and suitable application scenarios.
2
Urban planning, transformation, development and innovation needs
At present, we are facing the transformation of urban planning, and the previous planning emphasizes more on the planning of the scene and the static blueprint. However, in the planning stage, the people-oriented response is insufficient, and after entering the implementation stage, the means of dynamic monitoring mechanism need to be improved. The use of new technologies to empower planning, and the use of multiple data to support the preparation of planning, including spatial analysis and dynamic simulation, the construction of intelligent platforms, etc., have become the means to solve these problems.
At the same time, under the new background of land space and the new needs of modern governance, the smart planning process supported by multi-source big data generally has problems such as data inaccessibility and business coordination, and it is urgent to have a smart planning technical framework and application system based on multi-source data fusion to empower smart planning decision-making.
Facing the future, based on the background of the reconstruction of human-land relationship under the influence of ecological civilization and intelligent technology, big data and artificial intelligence technology are applied to all aspects of urban planning, and the urban intelligent planning framework integrating man-technology-space is constructed from the perspectives of human-environment system theory, spatio-temporal coupling, and spatial synthesis, which has become an important direction for planning transformation and innovative development.
3
AI-driven urban research and planning prospects
3.1 AI has brought new opportunities for the innovative development of urban research and planning
AI technologies are applied and integrated into the urban environment in different aspects of the city, forming urban AI (Luusua and Ylipulli 2020b), such as autonomous driving, robotics, and urban brains. They are ubiquitous, extending to spaces, places, and people's lives, and increasingly transforming the city into an intelligent living entity.
In addition to other sub-disciplines of urban science, urban planning will accelerate the cross-integration with computer, data science and other disciplines, and reconstruct the theoretical system of urban planning through "interconnection", which will further highlight its comprehensiveness, practicality and innovation, and enhance its scientificity. Artificial intelligence (AI) is rapidly emerging as a key technology that is transforming and reshaping the field of urban planning. However, there are still some unanswered questions about the potential impact of AI on urban and regional planning research and practice, the issues involved, and appropriate responses and plans (Peng, et al. 2023).
Complex urban problems can be designed with artificial intelligence or machine learning methods to design new models and new algorithms, which shows that new algorithms based on machine learning will open a new era of scientific computing for urban research and planning. Therefore, urban research and urban planning need to take a proactive and positive attitude and think about how to deal with the AI-driven urban research and planning and design paradigm from a new perspective.
3.2 AI-driven urban research and planning design: a paradigm shift
With the in-depth use of intelligent technology, people and urban space, technology and urban space are becoming more and more integrated, which makes urban physical space more and more intelligent.
The AI-driven scientific research paradigm is the deep integration of "artificial intelligence technology represented by machine learning" and "scientific research", which will promote the transformation of the data-driven urban research paradigm to the AI-driven urban research paradigm.
Urban planning AI is divided into AI-assisted, AI-enhanced, AI-automatic, and finally AI-autonomous planning (Peng, et al. 2003).
Future-proof, AI-driven paradigm shifts include:
1) Optimize the discipline system of urban science: including theory, method and application system, and improve the efficiency of scientific research;
2) Promote multidisciplinary integration: create a stable and high-quality scientific research ecology and cultivate innovative talents;
3) Build an urban open innovation platform: Based on artificial intelligence model and algorithm innovation, we will promote the construction of a multidisciplinary and multi-mode urban open innovation platform for major issues and key areas in the high-quality development of the city, from workshops to platforms;
4) Build multi-type comprehensive application scenarios for cities: strengthen the orientation of high-quality development of service cities.
3.3 AI-Driven Urban Research: Practical Applications
A city is a complex human-land system that includes social, economic, and ecological aspects. High-quality urban development needs to comprehensively consider the actual needs of scale and structure, function and efficiency, quality and vitality, safety and livability, etc., so as to realize the concept of green, low-carbon, intelligent, resilient and people-oriented development. Through certain policy, economic, technological and other regulation and control means, the rational allocation of factor resources is carried out, so as to optimize the spatial pattern and achieve high-quality development of the city.
By mapping the multi-dimensional characteristics, changes and operation process of urban land space to the virtual information space in a systematic and accurate and real-time manner, according to the different demand subjects and scenarios of the city, through the multi-state empowerment mode, the operation mechanism of the land space is understood, its organizational behavior is simulated, and intelligent early warning and management are carried out, so as to improve the intelligent response ability of the land space to complex problems, and make it a perceptible, thinking and adaptive living organism.
04
Conclusion
Forward-looking exploration and systematic design of AI-driven urban research is an inevitable trend in the innovative development of urban research and planning.
In the era of AI, urban research and planning should make new breakthroughs in the exploration of complex urban human-land relationships, so as to provide a theoretical basis and method for high-quality urban development.
The urban application of AI technology should not be viewed in isolation, but should be based on the integrated development of urban and rural planning, geography, ecological environment, computer and other sciences, and the impact of AI technology on the urban built environment, human, socio-economic space and information space and its coping strategies should be analyzed from the perspective of system integration.
Contributed by: Urban Planning New Technology Application Committee of Urban Planning Society of China
【Disclaimer】The content published on this official account is only for learning and communication, and is not for profit in any form. The copyright of the content belongs to the original author. If there is any infringement of your rights and interests, please contact us in time, and we will negotiate copyright issues or delete the content as soon as possible. The content is the author's personal opinion and does not represent the position of this official account and is responsible for its authenticity.
You might also want to watch these
Zhen Feng: Accelerate the construction of smart city agglomerations and help the high-quality development of the Yangtze River Delta
【Planning Annual Meeting】Zhen Feng: Thoughts on the Construction of Urban Intelligent Space Based on the Needs of People's Livelihood and Governance
Planning Field Problem | How to use smart technology to accurately assess and comprehensively monitor the operation status of the city and improve the urban physical examination system?