研究方向
· 复杂任务的多粒度知识建模
· 未知类别检测的开放集识别
· 新知识增量更新与连续学习
授课信息
· 数据挖掘
社会兼职
-
· 中国人工智能学会粒计算与知识发现专业委员会,副秘书长
-
· 天津市人工智能学会,理事
个人简介
王煜,博士,副教授,中国人工智能学会粒计算与知识发现专委会副秘书长,天津市人工智能学会理事;
入选天津市高层次青年人才(2024)、百大中国博士后科学基金资助者(2023);
主持国家自然科学基金面上项目、青年项目,科技部科技创新2030重大项目课题级任务,天津市重点研发计划科技支撑等重点课题等20余项科研项目;
围绕开放环境机器学习,发表CCF-A类/IEEE汇刊/SCI一区高水平论文近40篇,其中第一/通讯作者发表近30篇;
曾获得或指导学生获得中国国际大学生创新大赛金奖、CVPR开放世界挑战赛冠军、天津市优秀博士学位论文等10余项荣誉奖励;
研制技术已应用至现代医学与传统中医智能诊疗、关键设备故障诊断与健康管理、海洋牧场立体监测体系、社会治理等多个重要领域,技术入选工信部全国机器人应用场景优秀名单。
欢迎产学研合作及希望加入团队的硕博士研究生联系,请邮件咨询wang.yu@tju.edu.cn。
----- News ! -----
[2024/11] 课题组博士研究生姚鑫杰,硕士研究生季罗娜、穆郡贤获得2024年度研究生国家奖学金.
[2024/09] 论文"Persistence homology distillation for semi-supervised continual learning" 录用至CCF-A类会议 Annual Conference on Neural Information Processing Systems (NeurIPS).
[2024/09] 论文"What matters in graph class incremental learning? An information preservation perspective" 录用至CCF-A类会议 Annual Conference on Neural Information Processing Systems (NeurIPS).
[2024/08] 课题组团队获2024年“挑战杯”天津市大学生创业计划竞赛金奖.
[2024/07] 课题组硕士研究生谢嘉博获中国粒计算与知识发现学术会议最佳论文奖.
[2024/06] 论文"Boosting pseudo labeling with curriculum self-reflexion for attributed graph clustering" 录用至SCI一区TOP期刊IEEE Transactions on Neural Networks and Learning Systems (TNNLS).
[2024/05] 课题组博士研究生姚鑫杰获评天津大学智能与计算学部“杰出青年”.
[2024/05] 论文"Socialized learning: making each other better through multi-agent collaboration" 录用至CCF-A类会议 International Conference on Machine Learning (ICML).
[2024/04] 论文"Integrated heterogeneous graph attention network for incomplete multi-modal clustering" 发表至CCF-A类期刊International Journal of Computer Vision (IJCV).
[2024/02] 论文"Multi-view deep subspace clustering networks" 发表至SCI一区TOP期刊IEEE Transactions on Cybernetics (TCYB).
[2024/01] 论文"Exploring diverse representations for open set recognition" 录用至CCF-A类会议Proceedings of the AAAI Conference on Artificial Intelligence (AAAI).
[2024/01] 论文"Every node is different: dynamically fusing self-supervised tasks for attributed graph clustering"录用至CCF-A类会议Proceedings of the AAAI Conference on Artificial Intelligence (AAAI).
[2024/01] 论文"Dynamic sub-graph distillation for robust semi-supervised continual learning"录用至CCF-A类会议Proceedings of the AAAI Conference on Artificial Intelligence (AAAI).
[2023/12] 论文"Few-shot learning with multi-granularity knowledge fusion and decision-making"录用至国际权威期刊IEEE Transactions on Big Data (TBD).
[2023/12] 论文"Industrial big data analytical system in industrial cyber-physical systems based on coarse-to-fine deep network" 录用至国际权威期刊IEEE Transactions on Industrial Cyber-Physical Systems (TICPS).
[2023/07] 论文"Coarse-to-fine: progressive knowledge transfer based multi-task convolutional neural network for intelligent large-scale fault diagnosis"发表至SCI一区TOP期刊IEEE Transactions on Neural Networks and Learning Systems (TNNLS).
[2023/06] 课题组团队获CVPR 2023开放世界目标发现挑战赛冠军.
[2023/06] 课题组团队获CVPR 2023深度神经网络模型连续学习挑战赛亚军.
[2023/04] 论文"Class-specific semantic reconstruction for open set recognition"发表至CCF-A类期刊IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI).
[2023/04] 论文"Multi-granularity regularized re-balancing for class incremental learning"发表至CCF-A类期刊IEEE Transactions on Knowledge and Data Engineering (TKDE).
学术成果
论文成果
-
[1]Class-specific semantic reconstruction for open set recognition
-
[2]Integrated heterogeneous graph attention network for incomplete multi-modal clustering
-
[4]Multi-granularity regularized re-balancing for class incremental learning
-
[5]Persistence homology distillation for semi-supervised continual learning
-
[6]Exploring diverse representations for open set recognition
-
[7]Hierarchical semantic risk minimization for large-scale classification
-
[8]Deep fuzzy tree for large-scale hierarchical visual classification
-
[9]Boosting pseudo labeling with curriculum self-reflexion for attributed graph clustering
-
[11]Dynamic sub-graph distillation for robust semi-supervised continual learning
-
[12]Every node is different: dynamically fusing self-supervised tasks for attributed graph clustering
-
[13]Collaborative decision-reinforced self-supervision for attributed graph clustering
-
[14]Socialized learning: making each other better through multi-agent collaboration
-
[15]Latent heterogeneous graph network for incomplete multi-view learning
-
[17]Multi-granularity episodic contrastive learning for few-shot learning
-
[18]Uncertainty instructed multi-granularity decision for large-scale hierarchical classification
-
[20]大规模分类任务的分层学习方法综述