张得天 副教授

张得天，中国科学技术大学和香港城市大学计算机专业联合培养博士，苏州大学未来科学与工程学院副教授，硕士生导师。曾先后在香港城市大学、香港城市大学深圳研究院和江南大学担任过研究助理、副研究员和讲师等职位。主要研究兴趣为检索增强生成、图神经网络和时空大数据等。主持国家自然科学基金1项，参与国家级项目4项。以第一作者或通信作者先后发表了20余篇CCF A/B类或SCI 1/2区期刊和会议论文，并（曾）担任包括TKDE、TMC、TITS、IJCAI、AAAI、SIGKDD、CIKM和WSDM等在内的多个国内外知名期刊和会议的审稿人或PC member。

1、检索增强生成

2、图神经网络及应用

3、时空数据分析

相关方向近年部分论文如下：

期刊论文：

1. Haiyang Liu, Chunjiang Zhu, Detian Zhang*, Qing Li. Fully Adaptive Multi-Scale Spatial-Temporal Recurrent Networks for Traffic Flow Prediction.IEEE Transactions on Artificial Intelligence, Volume 7, Issue 4, pp 2088 - 2101, 2026.

2. Wenhe Xing, Jingsong Lv, Yipeng Zhou, Shiting Wen*, Detian Zhang*. Knowledge-Flow Contrastive Learning for Recommendation. Information Fusion, Volume 125, January 2026. (SCI-1区, CAAI-A)

3. Detian Zhang, Yunjun Zhou, Jin Wang*. Game Theory Based Approach for Massive Route Planning in Dynamic Road Networks. IEEE Transactions on Consumer Electronics (TCE),Volume 71, Issue 1, pp 2131-2141, 2025. (SCI-1区)

4. Detian Zhang, Hongwei Tang, Chunjiang Zhu*, Qing Li. Finding the Maximum Density Path withinConstrained Length in a Graph.IEEE Transactions on Intelligent Transportation Systems (TITS), Volume 26, Issue 6, pp 8513-8524, 2025. (SCI-1区)

5. Xunlian Luo, Chunjiang Zhu, Detian Zhang*, Qing Li. DSTAN: An Attention-Enhanced Dynamic Spatial-Temporal Network for Traffic Forecasting. World Wide Web Journal (WWWJ),Volume 28, Issue 15, pp 1-20, 2025. (CCF-B)

6. Shiting Wen, Yunjun Gao, Detian Zhang*, Jinqiu Yang, Qing Li. An Efficient Data Acquisition System for Large Numbers of Various Vehicle Terminals. IEEE Transactions on Intelligent Transportation Systems (TITS), Volume 23, Issue 5, pp 4720-4725, 2022. (SCI-1区)

7. Detian Zhang, Shiting Wen, Fei Chen, Zhixu Li, Lei Zhao. Spatial crowdsourcing based on Web mapping services. World Wide Web Journal (WWWJ), Volume 23, Issue 1, pp 631-648, 2020. (CCF-B)

会议论文：

1. Junwei Liu, Detian Zhang*.An Effective Distributed Ride-Sharing Framework on Large-Scale Road Networks. The 31st International Conference on Database Systems for Advanced Applications (DASFAA), Jeju, South Korea, 2026. (CCF-B)

2. Junwei Liu, Detian Zhang*. An Effective Multi-Hop Ride-Sharing Algorithm on Large-Scale Road Networks. The 26th International conference on web information systems engineering (WISE), Marrakech, Morocco, 2025. (CCF-B)

3. Detian Zhang, Chengqiang Zhang, Yanghui Rao, Qing Li, Chunjiang Zhu*. Improved Expressivity of Hypergraph Neural Networks through High-Dimensional Generalized Weisfeiler-Leman Algorithms. The 42nd International Conference on Machine Learning (ICML), Vancouver, Canada, 2025. (CCF-A)

4. Mingcong Shi, Chunjiang Zhu, Detian Zhang*, Shiting Wen, Qing Li. Multi-Granularity History and Entity Similarity Learning for Temporal Knowledge Graph Reasoning. The 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP), Miami, USA, 2024. (CCF-B)

5. Xunlian Luo, Chunjiang Zhu, Detian Zhang*, Qing Li. Dynamic Graph Convolutional Network with Attention Fusion for Traffic Flow Prediction. The 26th European Conference on Artificial Intelligence (ECAI), Kraków, Poland, 2023. (CCF-B)

6. Yunjun Zhou, Shuhan Wan, Detian Zhang*, Shiting Wen. Skilled Task Assignment with Extra Budget in Spatial Crowdsourcing. The 19th International Conference on Advanced Data Mining and Applications (ADMA), Shenyang, China, 2023. (CCF-C)

7. Haiyang Liu, Chunjiang Zhu, Detian Zhang*, Qing Li. Attention-based Spatial-Temporal Graph Convolutional Recurrent Networks for Traffic Forecasting. The 19th International Conference on Advanced Data Mining and Applications (ADMA), Shenyang, China, 2023. (CCF-C)

8. Detian Zhang, Lun Jin, Chunjiang Zhu*, Qing Li. Efficient Optimal Pick-up and Drop-off Point Recommendation for Ride-hailing Services. IEEE International Conference on Web Services (ICWS), Chicago, USA, 2023. (CCF-B)

9. Hongwei Tang, Detian Zhang*. Range Restricted Route Recommendation Based on Spatial Keyword. The 16thACM International Conference on Web Search and Data Mining (WSDM), Singapore, 2023. (CCF-B)

10. Lun Jin, Shuhan Wan, Detian Zhang*, Ying Tang. Extra-Budget Aware Online Task Assignment in Spatial Crowdsourcing. The 23nd International conference on web information systems engineering (WISE), Biarritz, France, 2022. (CCF-B)

11. Shuhan Wan, Detian Zhang*, An Liu, Junhua Fang. Extra-Budget Aware Task Assignment in Spatial Crowdsourcing. The 22nd International conference on web information systems engineering (WISE), Melbourne, Australia, 2021. (CCF-B)

完整版论文见：https://dblp.org/pid/18/9991.html

本小组招收硕士生和本科生，涉及如下方向，欢迎感兴趣同学联系：detian@suda.edu.cn 

1、检索增强生成（RAG）：检索增强生成是一种将外部知识检索与大模型相结合的前沿技术。它通过先从外部知识库或数据库中检索出与用户输入相关的信息，再将这些内容作为上下文输入到大模型中，从而生成更有依据和时效性的答案。该技术克服了传统大模型知识封闭、易编造事实的问题，广泛应用于问答系统、智能客服、搜索引擎增强等场景，特别适合需要实时查找和引用外部知识的任务。本小组研究涉及：RAG优化、RAG和知识图谱结合、多模态RAG、Agentic RAG、Agent Memory等。

2、图神经网络及应用：图神经网络是一类专门用于处理图结构数据的深度学习模型，能够有效捕捉节点之间的拓扑关系和特征依赖。其通过“信息传递”机制，在图中聚合邻居节点的信息，从而学习节点、边或整个图的表示。广泛应用于交通预测、知识图谱、社交网络分析、推荐系统、药物发现、金融风控等领域。本小组研究涉及：超图神经网络、知识图谱推理、基于图神经网络的推荐算法和交通流预测等。

3、时空数据分析：时空数据分析是对同时具有时间和空间属性的数据进行建模、处理和分析的一种方法，旨在揭示数据随时间和空间变化的内在规律。广泛应用于交通管理、物流调度、智慧城市、环境监测、公共安全等领域。本小组研究涉及：路径规划、交通流预测、k-NN、空间众包、司乘匹配和POI推荐等领域。