Research

A framework that constructs influence diagrams from natural-language descriptions using large language models, combining graph generation, probability elicitation, and iterative verification to support decision analysis.

Hong, Y., Qin, S.Z., Wang, C., 2026. LAMDA: Large Language Model as Decision Analyst. Decision Analysis, Articles in Advance. https://doi.org/10.1287/deca.2025.0438

@article{Qin20260807LAMDA,
  title = {LAMDA: Large Language Model as Decision Analyst},
  author = {Hong, Yifan and Qin, Sizhong and Wang, Chen},
  year = {2026},
  doi = {10.1287/deca.2025.0438},
  journal = {Decision Analysis}
}

Traceable language-model agents connect structural modeling, validation, analysis, checks, and reporting, evaluated with an executable benchmark.

Qin, S.Z., Gu, Y., Jiang, Y., Cai, A., Zhou, C.J., Shuai, S.X., Wang, J.C., Shen, T.H., Li, Y.Q., Li, X.H., Zeng, L., Chen, Y.S., Gao, D.C., Xu, G.R., Liao, W.J., Lu, X.Z. 2026. StructureClaw: Traceable LLM Agents and an Executable Benchmark for Structural Engineering Workflows. arXiv preprint arXiv:2607.14896. https://arxiv.org/abs/2607.14896

@misc{20260716StructureClaw,
  title = {StructureClaw: Traceable LLM Agents and an Executable Benchmark for Structural Engineering Workflows},
  author = {Qin, Sizhong and Gu, Yi and Jiang, Yao and Cai, Ao and Zhou, Changjian and Shuai, Shaoxuan and Wang, Jiachang and Shen, Tianhao and Li, Yueqiang and Li, Xinhao and Zeng, Li and Chen, Yueshi and Gao, Dachen and Xu, Genrong and Liao, Wenjie and Lu, Xinzheng},
  year = {2026},
  eprint = {2607.14896},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2607.14896}
}

An end-to-end advertisement video editing framework that uses multimodal discretization and controllable generation, unifying video selection, script generation, and BGM selection in a single pipeline.

Zhou, M., Qin, S.Z., Li, Y.Z., Chen, Q., Jiang, P., 2026. AutoCut: End-to-end advertisement video editing based on multimodal discretization and controllable generation. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 37777–37787. https://arxiv.org/abs/2603.28366

@inproceedings{20260330AutoCut,
  title = {AutoCut: End-to-end advertisement video editing based on multimodal discretization and controllable generation},
  author = {Zhou, Milton and Qin, Sizhong and Li, Yongzhi and Chen, Quan and Jiang, Peng},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2026},
  pages = {37777--37787}
}

A multimodal large language model that unifies floor plan understanding, generation, and editing via discrete room-instance tokens, enabling controllable and interpretable operations.

Qin, S.Z., Weber, R.E., Lu, X.Z., 2026. Tokenization Allows Multimodal Large Language Models to Understand, Generate and Edit Architectural Floor Plans. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp. 10430–10440. https://arxiv.org/abs/2603.11640

@inproceedings{20260316HouseMind,
  title = {Tokenization Allows Multimodal Large Language Models to Understand, Generate and Edit Architectural Floor Plans},
  author = {Qin, Sizhong and Weber, Ramon Elias and Lu, Xinzheng},
  booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2026},
  pages = {10430--10440}
}

This study proposes a diffusion model-based approach for predicting component dimensions in reinforced concrete frame structures, achieving high accuracy and engineering-consistent performance through multichannel masking and gradient-weighted correction.

Gu, Y., Qin, S.Z., Liao, W.J., Lu, X.Z., 2026. Intelligent design of dimensions of reinforced concrete frame structure components using diffusion models. Computers in Industry 175, 104428. DOI:10.1016/j.compind.2025.104428.

@article{Qin20260201CII,
  title = {Intelligent design of dimensions of reinforced concrete frame structure components using diffusion models},
  author = {Gu, Yi and Qin, Sizhong and Liao, Wenjie and Lu, Xinzheng},
  year = {2026},
  doi = {10.1016/j.compind.2025.104428},
  journal = {Computers in Industry},
  volume = {175},
  pages = {104428}
}

The study reviews data-driven intelligent optimization design for building structures, highlighting how data-driven AI can assist in initial solution generation, optimization problem simplification, optimization problem solving, and solution evaluation.

Qin, S.Z., Fei, Y.F., Liao, W.J., Lu, X.Z., 2025. Leveraging data-driven artificial intelligence in optimization design for building structures: A review. Engineering Structures 341, 120810. DOI:10.1016/j.engstruct.2025.120810.

@article{Qin20250624ES,
  title = {Leveraging data-driven artificial intelligence in optimization design for building structures: A review},
  author = {Qin, Sizhong and Fei, Yifan and Liao, Wenjie and Lu, Xinzheng},
  year = {2025},
  doi = {10.1016/j.engstruct.2025.120810},
  journal = {Engineering Structures},
  volume = {341},
  pages = {120810}
}

The study compares different algorithm combinations of BRB intelligent design and finds that GANs, diffusion models, and genetic algorithms perform well in capturing design features and exploring the solution space efficiently. Their integration achieves near-optimal and standard-compliant retrofit solutions, offering practical guidance for intelligent design applications.

Qin, S.Z., Liao, W.J., Tan, Z., Hu, K.G., Gao, Y., Lu, X.Z., 2025. Comparative analysis of intelligent retrofit design methods of RC frame structures using buckling-restrained braces. Bulletin of Earthquake Engineering 23(8), 3353–3374. DOI:10.1007/s10518-025-02164-3.

@article{Qin20250415BEE,
  title = {Comparative analysis of intelligent retrofit design methods of RC frame structures using buckling-restrained braces},
  author = {Qin, Sizhong and Liao, Wenjie and Tan, Zhuang and Hu, Kongguo and Gao, Yuan and Lu, Xinzheng},
  year = {2025},
  doi = {10.1007/s10518-025-02164-3},
  journal = {Bulletin of Earthquake Engineering},
  volume = {23},
  number = {8},
  pages = {3353-3374}
}

Presents a novel GNN-assisted evolutionary algorithm that integrates a powerful GNN surrogate model with evolutionary algorithms to rapidly optimize RC shear wall structures, significantly reducing computational costs while achieving high-quality design outcomes.

Fei, Y.F., Qin, S.Z., Liao, W.J., Guan, H., Lu, X.Z., 2025. Graph neural network-assisted evolutionary algorithm for rapid optimization design of shear-wall structures. Advanced Engineering Informatics 65, 103129. DOI:10.1016/j.aei.2025.103129.

@article{Qin20250115ADVEI,
  title = {Graph neural network-assisted evolutionary algorithm for rapid optimization design of shear-wall structures},
  author = {Fei, Yifan and Qin, Sizhong and Liao, Wenjie and Guan, Hong and Lu, Xinzheng},
  year = {2025},
  doi = {10.1016/j.aei.2025.103129},
  journal = {Advanced Engineering Informatics},
  volume = {65},
  pages = {103129}
}

Proposes a general heterogeneous graph representation method for RC frame structures capturing intrinsic properties and topological relationships. Develops an efficient method using HetGNN to determine the sectional sizes of RC frame structures within one second.

Qin, S.Z., Liao, W.J., Huang, Y.L., Zhang, S.L., Gu, Y., Han, J., Lu, X.Z., 2025. Intelligent design for component size generation in reinforced concrete frame structures using heterogeneous graph neural networks. Automation in Construction 171, 105967. DOI:10.1016/j.autcon.2025.105967.

@article{Qin20250113AIC,
  title = {Intelligent design for component size generation in reinforced concrete frame structures using heterogeneous graph neural networks},
  author = {Qin, Sizhong and Liao, Wenjie and Huang, Yuli and Zhang, Shulu and Gu, Yi and Han, Jin and Lu, Xinzheng},
  year = {2025},
  doi = {10.1016/j.autcon.2025.105967},
  journal = {Automation in Construction},
  volume = {171},
  pages = {105967}
}

Proposes a two-stage intelligent design method for the retrofitting of RC frame structures with BRBs, based on generative AI and optimization algorithms. This method enables the decoupling of architectural and structural design requirements.

Tan, Z., Qin, S.Z., Hu, K.G., Liao, W.J., Gao, Y., Lu, X.Z., 2025. Intelligent generation and optimization method for the retrofit design of RC frame structures using buckling-restrained braces. Earthquake Engineering & Structural Dynamics 54(2), 530–547. DOI:10.1002/eqe.4268.

@article{Qin20241114EESD,
  title = {Intelligent generation and optimization method for the retrofit design of RC frame structures using buckling‐restrained braces},
  author = {Tan, Zhuang and Qin, Sizhong and Hu, Kongguo and Liao, Wenjie and Gao, Yuan and Lu, Xinzheng},
  year = {2025},
  doi = {10.1002/eqe.4268},
  journal = {Earthquake Engineering \& Structural Dynamics},
  volume = {54},
  number = {2},
  pages = {530-547}
}

Develops a novel approach that utilizes large language models, graphormer, and diffusion models to generate and edit floor plans interactively.

Qin, S.Z., He, C.Y., Chen, Q.Y., Yang, S., Liao, W.J., Gu, Y., Lu, X.Z., 2024. ChatHouseDiffusion: Prompt-Guided Generation and Editing of Floor Plans. arXiv preprint arXiv:2410.11908. https://arxiv.org/abs/2410.11908

Develops an intelligent design and optimization system for shear wall structures, increasing efficiency by about 30 times.

Qin, S.Z., Guan, H., Liao, W.J., Gu, Y., Zheng, Z., Xue, H.J., Lu, X.Z., 2024. Intelligent design and optimization system for shear wall structures based on large language models and generative artificial intelligence. Journal of Building Engineering 95, 109996. DOI:10.1016/j.jobe.2024.109996.

@article{Qin20240701JOBE,
  title = {Intelligent design and optimization system for shear wall structures based on large language models and generative artificial intelligence},
  author = {Qin, Sizhong and Guan, Hong and Liao, Wenjie and Gu, Yi and Zheng, Zhe and Xue, Hongjing and Lu, Xinzheng},
  year = {2024},
  doi = {10.1016/j.jobe.2024.109996},
  journal = {Journal of Building Engineering},
  volume = {95},
  pages = {109996}
}

Constructs a local–cloud collaborative mode, introduces a comprehensive data transmission format, and develops a cloud interface for generative AI algorithms.

Qin, S.Z., Liao, W.J., Huang, S.N., Hu, K.G., Tan, Z., Gao, Y., Lu, X.Z., 2024. AIstructure-Copilot: Assistant for generative AI-driven intelligent design of building structures. Smart Construction 1(1), 0001. DOI:10.55092/sc20240001.

@article{Qin20240304SC,
  title = {AIstructure-Copilot: Assistant for Generative AI-Driven Intelligent Design of Building Structures},
  author = {Qin, Sizhong and Liao, Wenjie and Huang, Shengnan and Hu, Kongguo and Tan, Zhuang and Gao, Yuan and Lu, Xinzheng},
  year = {2024},
  doi = {10.55092/sc20240001},
  journal = {Smart Construction}
}

(in Chinese) The point cloud semantic segmentation using deep learning is introduced to realize the object-oriented comparison of components for the frame nodes, which increases the automation and digitization level of point-cloud-vs-BIM comparison.

Zhang, F., Sun, C.J., Qin, S.Z., Zhao, X.Y., 2024. Synthetic research on construction inspection based on BIM and point cloud segmentation using deep learning. Engineering Mechanics 41(2), 194–201. DOI:10.6052/j.issn.1000-4750.2022.04.0281.

This study introduces an assessment method used in the intelligent design and optimization of shear wall structures that effectively combines mechanical analysis and formulaic encoding of empirical rules.

Qin, S.Z., Liao, W.J., Lin, Y.Q., Lu, X.Z., 2023. An efficient assessment method for intelligent design results of shear wall structure based on mechanical performance, material consumption, and empirical rules. Engineering Mechanics 40(12), 148–159. DOI:10.6052/j.issn.1000-4750.2023.05.0360.

(in Chinese) This paper proposes an application framework for large language models in construction engineering, utilizing prompt engineering and a local knowledge base to enhance model performance.

Qin, S.Z., Zheng, Z., Gu, Y., Lu, X.Z., 2023. Exploring and discussion on the application of large language models in construction engineering. Industrial Construction 53(9), 162–169. DOI:10.13204/j.gyjzG23081006.

(in Chinese) Implications of the 7.8-magnitude earthquake in Turkey for Chinese earthquake prevention and mitigation efforts

Lu, X.Z., Qin, S.Z., Xu, Z., 2023. Enlightenment to China on earthquake prevention and disaster reduction from the M7.8 earthquake in Turkey. City and Disaster Reduction 2023(2), 1–8. DOI:10.3969/j.issn.1671-0495.2023.02.002.

(in Chinese) This study proposes a convenient calculation system for prefabricated segmental beam template adjustment based on the "cloud+end" model

Liu, Y., Ma, Z.L., Liu, S.L., Qin, S.Z., Zhou, F.N., Shi, Q.B., 2022. A calculation system for the adjustment of prefabricated segmental beam templates based on the "cloud + end" model. In: Proceedings of the 8th National Conference on Building Information Modeling, pp. 109–113. China Architecture & Building Press.

This paper proposes a quick location system for formwork adjustment based on a smartphone and the latest technologies including computer vision, BIM, and cloud computing.

Liu, Y., Ma, Z.L., Qin, S.Z., Liu, S.L., 2022. A Cloud-Based System for Supporting Quick Location of Precast Concrete Formwork Using Computer Vision and BIM. In: Proceedings of the Creative Construction e-Conference 2022, pp. 97–104. DOI:10.3311/CCC2022-014.

@inproceedings{Qin20220709CCC,
  title = {A Cloud-Based System for Supporting Quick Location of Precast Concrete Formwork Using Computer Vision and BIM},
  author = {Liu, Yu and Ma, Zhiliang and Qin, Sizhong and Liu, Shilong},
  year = {2022},
  doi = {10.3311/ccc2022-014},
  pages = {97-104},
  booktitle = {Proceedings of the Creative Construction e-Conference 2022}
}