AI和学科交叉背景下风景园林自主性的潜在路径：知识图谱+循证设计+寻解研究

Potential Pathways for Autonomy of Landscape Architecture within the Context of AI and Interdisciplinarity: Knowledge Graph + Evidence-Based Design + Solution-Oriented Research

关键词:  风景园林  循证设计  交叉学科  知识图谱  寻解研究

Key words:  landscape architecture  evidence-based design  interdisciplinary study  knowledge graph  practice-oriented research

摘要:

多学科交叉与人工智能(AI)的迅速发展对风景园林学科既是挑战也是机遇。以知识与实践的关系演变为背景，探索跨学科语境下风景园林知识边界的坚守与协同，提出AI技术支撑下“知识图谱-循证设计-寻解研究”三位一体逻辑链，以推进行业自主性与价值体系重构。知识图谱作为核心，通过数据整合与本体建模，应对知识碎片化与异构化问题，实现跨学科知识转译应用。循证设计借鉴医学“诊断-方案-实施-随访”的诊疗逻辑，以证据支撑设计决策。寻解研究从实践需求出发，遵循“需求-问题-结论-策略”路径，强化实践需求和问题导向与实践应用的结合。建议行业优先推进寻解研究，逐步制度化循证设计，分领域推进知识图谱建设。

Abstract:

The rapid development of artificial intelligence (AI) and accelerating multidisciplinary integration presents a critical juncture for contemporary landscape architecture, confronting the discipline with profound challenges such as increasingly blurred disciplinary boundaries, persistently fragmented knowledge systems, and a persistent disconnect between research and practical application, while simultaneously offering unprecedented opportunities to reconstruct its intellectual autonomy and professional value system through strategic technological empowerment. This paper, systematically examining the evolving "knowledge-practice" relationship against its historical backdrop, proposes an integrated theoretical framework of "Knowledge Graph + Evidence-Based Design + Solution-Oriented Research". This comprehensive framework aims to catalyze a fundamental paradigm shift from traditional experience-driven approaches towards rigorously evidence-based and systematically data-driven methodologies, thereby strategically fostering the discipline's autonomous development and ensuring its long-term sustainability within the rapidly evolving technological landscape. The article begins by tracing the historical evolution of the relationship between knowledge and practice in the human settlement environment, delineating four distinct phases: the ancient "Experience Symbiosis Period", where knowledge was embedded within craft and tradition; the modern "Scientific Separation Period", which saw the dominance of explicit scientific knowledge and a growing rift with practice; the postmodern "Interdisciplinary Transition Period," characterized by loosening disciplinary boundaries and the beginnings of knowledge integration; and the current "Technological Coupling Period", where AI and digital technologies are profoundly reshaping the production, integration, and application of knowledge. Within this context, the paper argues that AI technologies, particularly the synergy between knowledge graphs and large language models, offer novel possibilities for addressing persistent issues in landscape architecture, such as knowledge heterogeneity and cognitive overload. The core logical framework positions the "Knowledge Graph" as the central nervous system of the discipline's knowledge infrastructure. Through processes of knowledge acquisition, fusion, and dynamic updating, it facilitates the semantic alignment and structured storage of interdisciplinary knowledge, thereby enabling efficient articulation between research and practice. The knowledge graph integrates not only explicit scientific findings but also, through a rigorous screening mechanism, validated tacit practical experience, forming a shared, iterative "knowledge base" for the profession. At the practice level, "Evidence-Based Design" is conceptualized through an analogy to medical diagnostic logic, following a closed-loop process of "Examination and Diagnosis - Design Proposal - Implementation - Performance Follow-up". This approach emphasizes that design decisions should be grounded in quantifiable, verifiable evidence, thereby enhancing the functionality, ecological performance, and socio-health benefits of landscape projects. Even in the absence of a fully developed knowledge graph, the application of evidence in the design process can be advanced through systematic literature reviews and the distillation of design guidelines. On the research front, "Solution-Oriented Research" establishes a pathway that starts from practical problems, adhering to the chain of "Practical Need - Research Question - Research Conclusion - Solution Strategy". This ensures that research outcomes can directly inform and feed back into planning, design, and management practices. This model underscores the "usefulness" of research and advocates for researchers to balance breadth and depth by focusing on specific types of sites, thereby avoiding the pitfall of being "comprehensive yet superficial". Regarding implementation strategy, the paper proposes a phased approach: the profession should prioritize promoting "Solution-Oriented Research" to strengthen the problem-orientation and practical value of academic inquiry; subsequently, it should gradually institutionalize the processes of "Evidence-Based Design", focusing systematic reviews and guideline development on key domains; ultimately, the construction of the "Knowledge Graph" should be advanced step-by-step across various sub-fields, addressing the technical challenges of data integration, ontological modeling, and multi-modal knowledge fusion. In conclusion, the paper contends that the synergistic advancement of the "Knowledge Graph + Evidence-Based Design + Solution-Oriented Research" closed-loop logic enables landscape architecture to achieve the goals of "making design scientific enough", "ensuring research is distinctive", and "effectively applying knowledge from other disciplines". This, in turn, establishes the discipline's irreplaceable autonomy and self-evident value within the era of interdisciplinary convergence and AI integration. This framework not only provides theoretical support for the construction and translation of the profession's knowledge system but also charts a practical path for its future innovation empowered by technology.