个人资料:
姓名:田宏硕
职称:助理研究员
学科专业:信息与通信工程
通讯地址:bet365官网电气bet365官网与信息工程学院26教学楼D区
电子信箱:kellyeden@0313daikuan.com
主要经历:
(1) 2023.06-至今 bet365官网 电气bet365官网与信息工程学院 电子信息工程系,博士后/助理研究员
(2) 2018.09-2023.06 bet365官网,信息与通信工程专业,工学博士
(3) 2014.09-2018.06 bet365官网,电子信息工程专业,工学学士
主要研究方向:
(1) 多模态知识图谱: 视觉场景图生成
(2) 跨媒体理解:“视觉-语言”相互生成
(3) 视觉问答/对话
主要科研项目:
(1) 2022.01-2025.12 国家联合基金重点,参与
(2) 2021.01-2023.12 国家重点研发项目,参与
(3) 2021.01-2021.12 天津市研究生科研创新项目:“基于深度学习的多源域域适应算法研究”,第二负责人
代表性论著、学术著作:
学术论文:
(1) An-An Liu, Hongshuo Tian et al. Toward Region-Aware Attention Learning for Scene Graph Generation. IEEE Transactions on Neural Networks and Learning Systems. 2021 (SCI 一区)
(2) Hongshuo Tian, Ning Xu et al. Part-Aware Interactive Learning for Scene Graph Generation. ACM Multimedia 2020. (CCF A类会议)
(3) Hongshuo Tian, Ning Xu et al. Mask and Predict: Multi-Step Reasoning for Scene Graph Generation. ACM Multimedia 2021. (CCF A类会议)
(4) Hongshuo Tian, Ning Xu et al. Towards Confidence-Aware Commonsense Knowledge Integration for Scene Graph Generation. IEEE International Conference on Multimedia & Expo (ICME) 2023. (CCF B类会议)
(5) Ning Xu, Hongshuo Tian et al. Coupled-dynamic learning for vision and language: Exploring Interaction between different tasks. Pattern Recognition. 2021 (SCI 一区)
(6) Yanhui Wang, Ning Xu, Hongshuo Tian et al. Knowledge Prompt Makes Composed Pre-Trained Models Zero-Shot News Captioner. IEEE International Conference on Multimedia & Expo (ICME) 2023. (CCF B类会议)
(7) An-An Liu, Chenxi Huang, Ning Xu, Hongshuo Tian et al. Counterfactual Visual Dialog: Robust Commonsense Knowledge Learning from Unbiased Training. IEEE Transactions on Multimedia. 2023 (SCI 一区)
专利:
(1) Visual Relationship detection method and system based on region-aware learning mechanisms. (Patent No.: US 11301725)(已授权的专利)
(2) 一种基于区域注意力学习机制的视觉关系检测方法,发明专利,专利号:CN202010079031.6(已授权的专利)
其他(社会兼职等):
(1) 担任TCYB, TNNLS, TMM, TCSVT, ToMM, CVPR, AAAI, ICLR, ICML, ACM MM,等国际顶级期刊与会议审稿人