建筑业人工智能研究主题识别及演化分析

韩映雪

摘 要

随着近几年人工智能被国家列为产业融合发展和转型升级核心技术，建筑业作为国民支柱产业，利用人工智能作为推进数字化转型的关键技术，是推动建筑行业高质量发展的关键。跟随国家政策指引，国内外对人工智能相关研究数量也在飞速攀升，梳理国内建筑业人工智能研究主题，为前沿研究探明研究方向和路径，是推动建筑业人工智能快速发展重要保障。

本文通过对1985-2021年国内外建筑业人工智能相关研究文献进行梳理，将隐含狄利克雷分布（Latent Dirichlet Allocation，LDA）引入建筑业人工智能前沿分析研究中，挖掘出国内外主题构成。再引入时间变量，划分时间窗口，对主题强度和主题内容进行演化分析，之后综合主题强度和主题内容演化结果对国内外建筑业人工智能研究阶段进行划分，并对国内外主题研究进行对比分析。最后构建多维指标对国内建筑业人工智能主题进行前沿预测，并提出建筑业人工智能发展针对性政策建议。

研究得出国内建筑业人工智能行业发展首先需要加强人工智能底层技术的研究，构建技术人才培养和发展的长效机制。其次促进建筑业人工智能环境优化、数字技术、建筑空间优化等应用和底层技术的深度融合。此外还需推动建筑业人工智能从多个层面与国外开发合作，加强相关理论和关键技术的研究。最后需要持续关注计算机视觉、建筑灾害事故、建筑施工环境优化、建筑业全生命周期信息化和数据信息安全等当前热点主题研究，重点关注以机器学习等前沿趋势主题研究，稳步推进城市空间优化和建筑文化遗产保护正在探索的主题研究，逐步落实建筑三维可视化、行业数字化教育、虚拟现实技术等成熟区主题研究，使研究资源向相关领域倾斜，助力建筑业人工智能研究发展。

关键词：建筑业；人工智能；主题演化；LDA主题模型；前沿识别

Abstract

In recent years, artificial intelligence has been listed as the core technology of industrial integration development and transformation and upgrading by the country. As a national pillar industry, the utilization of ARTIFICIAL intelligence as a key technology to promote digital transformation is the key to promote high-quality development of the construction industry. Following the guidance of national policies, the number of artificial intelligence-related researches at home and abroad is also rising rapidly. Sorting out the research topics of artificial intelligence in the domestic construction industry is an important guarantee to promote the rapid development of artificial intelligence in the construction industry, so as to clarify the research direction and path for frontier research.

In this paper, the LDA theme model is introduced into the frontier analysis and research of artificial intelligence in the construction industry by combing the domestic and foreign literature related to artificial intelligence from 1985 to 2021, and the theme composition at home and abroad is excavated. Then the time variable is introduced to divide the time window, and the evolution of the topic intensity and the topic content is analyzed. After that, the research stages of artificial intelligence in the construction industry at home and abroad are divided based on the evolution results of the topic intensity and the topic content, and the comparative analysis of the topic research at home and abroad is conducted. Finally, a multi-dimensional index is constructed to forecast the artificial intelligence theme in the domestic construction industry and put forward targeted policy suggestions for the development of artificial intelligence in the construction industry.

It is concluded that the development of artificial intelligence industry in domestic construction industry first needs to strengthen the research on the underlying technology of artificial intelligence, and build a long-term mechanism for the cultivation and development of technical personnel. Secondly, the application of artificial intelligence environment optimization, digital technology and building space optimization in construction industry should be deeply integrated with the underlying technology to achieve technology landing. In addition, it is necessary to promote the construction industry AI cooperation with foreign development from multiple levels, focus on global wisdom, and strengthen the research of relevant theories and key technologies. The need to continue to focus on computer vision, construction disasters and accidents, the construction environment optimization, the whole life cycle of construction information and data information security such as the current hot topic research, focus on the study theme is given priority to with machine learning the forefront of the trend, steadily push forward the urban space optimization and architectural culture heritage protection are exploring the theme of the research, Gradually implement the subject research of mature areas such as 3d architecture visualization, industry digital education and virtual reality technology, tilt the research resources towards related fields, and help the research and development of artificial intelligence in construction industry.

Key words: The construction industry; Artificial intelligence; Topic evolution; LDA topic model; Cutting edge recognition