姓名:徐菲菲

性别:女

职称:教授

Email:xufeifei@shiep.edu.cn

办公地址:人工智能学部A308

个人简介:

在科学研究方面，主持国家自然科学基金1项，主持上海市“晨光人才”计划1项，主持上海市教委科研创新、优青项目各1项；申请发明专利13项，其中已授权8项。作为核心研究人员参与国家自然科学基金面上项目4项，省部级项目9项，总经费超过400万元。近五年来，以第一作者发表学术论文40余篇，其中CCF 顶会论文5篇，SCI检索论文4篇，中科院一、二区论文3篇，CCF C类论文26篇。获CCF顶会DASFAA2022唯一整体最佳论文奖、北大核心期刊《数据采集与处理》2019-2021年度优秀论文奖，获上海电力大学智能电网研究院“特聘研究员”。

在教育教学方面，主持全国计算机基础协会教学改革项目1项，主持上海市“课程思政重点改革领航学院”课程思政项目1项。指导学生获得国家、市级学科竞赛荣誉16项，指导学生市级科创项目4项。多次获得“光力科技奖教金”、“我心目中的好老师”、青年教师讲课比赛二等奖等荣誉。

在硕士培养方面，共培养毕业研究生26名。其中5名获国家奖学金，1名获上海市优秀毕业生，8名获校优秀毕业生。荣获校“优秀党务工作者”、“三八红旗手”、“征兵先进个人”等荣誉称号及“上海市优秀青年女教师成才资助金”等。

研究方向:
人工智能、自然语言处理、多模态、大模型、多智能体

主讲课程:
程序设计基础、数据库原理、数据挖掘

科研成果:
[1] Feifei Xu, Wang Zhou, Fumiaoyue Jia. Multimodal Prompting and   Masking strategy for Video-Grounded Dialogue.KNOWLEDGE-BASED   SYSTEMS,2025(SCI，中科院一区，影响因子：7.6)

[2] Feifei Xu,Ziheng Yu. Navigating Style Variations in Scene Text   Image Super-Resolution through Multi-Scale Perception.ICMR ＇24:   Proceedings of the 2024 International Conference on Multimedia Retrieval(CCF   B，EI)

[3] Feifei Xu,Qinghan Du, Xiaoyan Yu.VF-Dual:   View-Fusion Dual-Stream Architecture for Emotion Recognition in   Conversations.International Conference on Intelligent Computing,   2025(CCF C,EI)

[4] Feifei Xu,Qinghan Du,Luobin Huang. Multi-View Decoupling   Framework for Emotion Recognition in Conversations. IEEE International   Conference on Ubiquitous Intelligence and Computing，   2024(CCF C,EI)

[5] Feifei Xu,Haoran Bi, Bo Li.ProCFD:   Towards Robust Multimodal Sentiment Analysis through Prototype Fusion and   Contrastive Feature Decomposition.International Conference on Intelligent   Computing, 2025(CCF C,EI)

[6] Feifei Xu,Qiye Cai, Fumiaoyue Jia. Enhancing CLIP for   Pedestrian Image-Text Retrieval via Bi-level Alignment and Weighted   Similarity Distribution Matching Loss.International   Conference on Intelligent Computing, 2025(CCF C,EI)

[7] Feifei Xu,Tao Sun, Wang   Zhou. Multi-Level Causal Reasoning for Emotion Recognition in Conversations.The 21st IEEE International   Conference on Ubiquitous Intelligence and Computing (UIC 2024) (CCF C,EI)

[8] Feifei Xu, Fumiaoyue   Jia, Wang Zhou.Multimodal Prompt Learning for Audio Visual Scene-Aware   Dialog.31st   International Conference on Multimedia Modeling, MMM 2025(CCF C,EI)

[9] Feifei Xu, Jiahao Lu, Haizhou Du.Context Dynamics   Tracking and Curriculum Learning for Emotion Recognition in Conversation.The   11th International Conference on Behavioural and Social Computing,2024(EI)

[10] Feifei Xu,Yingchen   Zhou, Zheng Zhong. Object Category-Based Visual Dialog for Effective Question   Generation. International   Conference on Computational Visual Media, 2024(CCF C,EI)

[11] Feifei Xu,Yingchen Zhou, Zheng Zhong. Enhancing Relevance and   Efficiency in Visual Question Generation Through Redundant Object Filtering.10th   International Conference of Pioneering Computer Scientists, Engineers and   Educators, ICPCSEE 2024(EI)

[12] Feifei Xu,Tao Sun, Wang Zhou CTF-ERC: Coarse-to-Fine   Reasoning for Emotion Recognition in Conversations. the International Joint   Conference on Neural Networks, 2024(CCF C, EI)

[13] Feifei Xu,Zheng Zhong,Yitao Zhu.Object-based   Appearance-Motion Heterogeneous Network for Video Question Answering.2023   IEEE 29th International Conference on Parallel and Distributed Systems   (ICPADS) (CCF C, EI)

[14] Feifei Xu,Zheng   Zhong,Yitao Zhu. Appearance-Motion Dual-stream Heterogeneous Network for   VideoQA.International Conference on Multimedia Modeling.2024(CCF C, EI)

[15] Feifei Xu, Wang   Zhou, Guangzhen Li.Enhancing Cross-Modal Understanding for Audio Visual   Scene-Aware Dialog Through Contrastive Learning.2024 IEEE International Symposium on Circuits and   Systems (ISCAS) (CCF C, EI)

[16] Feifei Xu,Chun   Wang, Yitao Zhu. Emoji-Inserted Emotional Dialogue Generation with   Emoji-Enriched Knowledge Graph.16th   International Conference on Advanced Computer Theory and Engineering (ICACTE)   (CCF C, EI)

[17] Feifei Xu, Wang Zhou,Tao   Sun.A Coarse and Fine Grained Masking Approach for Video-grounded Dialogue.30th International   Conference on Multimedia Modeling (MMM 2024) (CCF C, EI)

[18] Feifei Xu,Guangzhen   Li, Zheng Zhong.D-Man: A Distance-Based Multi-Channel Attention   Network for ERC.2024   International Joint Conference on Neural Networks (IJCNN) (CCF C, EI)

[19] Feifei Xu, Shanlin Zhou, Xinpeng Wang, Wenkai Zhang.Apeak-CG:   Automatically predicting emotion based dynamic multi-form knowledge fusion   conversation generation. Neurocomputing, Volume 504, 14 September 2022, Pages   174-186(SCI, JCR一区，中科院二区，影响因子：5.779)

[20] Feifei Xu, Shanlin Zhou, Yunpu Ma, Xinpeng Wang, Wenkai   Zhang, Zhisong Li. Open-domain Dialogue Generation Grounded with Dynamic   Multi-form Knowledge Fusion. Database Systems for Advanced Applications   2022(DASFAA2022), Hyderabad, India: 101-116 (CCF-B类会议,   overall best paper,EI).

[21] Feifei Xu, Xinpeng Wang, Yunpu Ma, Volker Tresp, Yuyi Wang,   Shanlin Zhou, Haizhou Du. Controllable Multi-Character Psychology-Oriented   Story Generation. The 29th ACM International Conference on Information and   Knowledge Management(CIKM2020), Ireland: 1675–1684(CCF-B类会议,EI)

[22] Feifei Xu, Xinpeng Wang, Shanlin Zhou. Story Generation Using   Knowledge Graph under Psychological States, Wireless Communications and   Mobile Computing, 2021(1): 1-12. (SCI, CCF C类，影响影子：2.336) 

[23] Feifei Xu, Wenkai Zhang, Haizhou Du, et al. Enhancing   Machine Comprehension Using Multi-Knowledge Bases and Offline Answer Span   Improving System. Journal of Internet Technology, 2021,   22(5):1095-1107. DOI:10.53106/160792642021092205013 (SCI，影响因子：1.14)

[24] Feifei Xu, Guangzhen Li, et al. Multiple Information   Extraction and Interaction for Emotion Recognition in Multi-Party   Conversation. The 28th IEEE Symposium on Computers and   Communications(ISCC2023), Tunis, Tunisia, 2023-7-9(CCF-C，EI)

专利：

[1]徐菲菲,王新鹏,周珊琳. 一种基于深度学习的可控多人物心理导向故事生成方法,2022-6-21,中国, ZL202011071294.9 (已授权，发明专利)

[2]徐菲菲,张文楷.一种基于知识图谱增益的机器阅读理解模型,2022-8-9,中国,ZL202110082108.X (已授权，发明专利)

[3]徐菲菲,刘晶晶.基于深度学习的轻量级和多姿态人脸检测与识别系统,2021-7-16,中国,ZL202022623118.3（已授权，实用新型）

[4]徐菲菲,周珊琳.基于动态多形式知识融合的开放域对话生成方法,中国,CN202211549074.1(已授权，发明专利)

[5]徐菲菲,王纯.利用融入表情符号的常识知识图谱优化情感回复的方法,中国(已授权，发明专利)

[6]徐菲菲,王纯.基于深度学习的带有表情符号的情感回复生成方法,中国(已授权，发明专利)

[7] 徐菲菲,田宇.一种基于稠密网络和多相似损失的服饰检索算法模型,中国(申请中，发明专利)

[8]徐菲菲,田宇.一种基于孪生网络和稠密网络的服装检索模型,中国(申请中，发明专利)

[9]徐菲菲,曹阳泽.基于知识图谱的语义感知自动文本摘要方法,中国(已授权，发明专利)

[10]徐菲菲,曹阳泽.基于多知识图谱的适配主题优选文本摘要方法,中国(已授权，发明专利)

科研项目:
[1] 国家自然科学基金青年项目.基于粒计算与语义模板的问答系统研究. 2014.1-2016.12.主持，已结题

[2] 晨光计划.基于粒计算的智能电网负荷预测关键技术.2014.7-2017.6.主持，已结题

[3] 上海市教委科研创新计划.基于云计算的粗糙集属性约简方法研究及其在电力系统的应用.2012.1-2014.12.主持，已结题

[4] 上海市教委优青项目.基于模糊粗糙集的大型燃煤机组煤耗关键因素研究. 2012.1-2013.12.主持，已结题