
吕新昱
新财经综合实验室
副教授、硕士生导师
个人主页:https://xinyulyu.github.io/
E-mail:xinyulyu68@gmail.com
yl23411永利官网登录柳林校区经世楼D205
※教师简介
吕新昱,副教授,硕士生导师,现任职于永利yl23411集团官网(新财经综合实验室)。2024年6月毕业于电子科技大学计算机科学与工程学院,博士导师为高联丽教授、宋井宽教授与邵杰教授,于2024年6月获得工学博士学位。主要研究方向为大模型安全与多模态处理。近五年已完成20余篇高水平学术论文,其中12篇发表于CCF-A类会议或中科院SCI-1区Top期刊,包括IEEE TPAMI、IJCV、IEEE TIP、IEEE TCSVT、PR、NeurIPS、CVPR、ACMMM等。主持四川省青年基金1项,参与国家自然科学基金项目3项、科技部科技创新2030"新一代人工智能"重大项目1项,以及快手高校科研项目2项等。目前担任多个国际领域会议和顶级期刊审稿人如IEEE TPAMI、IEEE TMM、ACM TOMM、CVPR、ICCV、ECCV、AAAI和ACM MM等。2023年获得四川省计算机学会SCF优秀论文奖。
欢迎对大模型安全、多模态处理等方向感兴趣的同学发送简历至我的邮箱,我会尽快回复邮件并安排面谈。
※研究领域
大模型安全,多模态处理,计算机视觉
※教育背景
2020/09-2024/07 电子科技大学 博士
2017/09-2019/07 美国罗格斯大学 硕士
2014/09-2018/07 电子科技大学 本科
※职业经历
2024/07年至今 yl23411永利官网登录计算机与人工智能学院 副教授
※研究成果
代表性学术论文(*通讯作者)
[1] Lyu, Xinyu and Gao, Lianli* and Zeng, Pengpeng and Shen, Heng Tao and Song, Jingkuan. Adaptive Fine-Grained Predicates Learning for Scene Graph Generation, IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023: 13921-13940, (中科院一区Top期刊,影响因子:23.6,财大外文A+).
[2] Lyu, Xinyu and Gao, Lianli* and Xie Junlin, and Zeng, Pengpeng and Tian, Yulu and Shao, Jie and Shen, Heng Tao. Multi-Concept Learning for Scene Graph Generation. IEEE Transactions on Image Processing, 2025. (中科院一区Top期刊,影响因子:11.6,财大外文A).
[4] Lyu, Xinyu and Chen, Beitao and Gao, Lianli and Song, Jingkuan and Shen, Heng Tao. Alleviating Hallucinations in Large Vision-Language Models through Hallucination-Induced Optimization, Proceedings of the 38th Conference on Neural Information Processing Systems (NeurIPS), 2024, (机器学习领域顶级会议,CCF A).
[3] Lyu, Xinyu and Gao, Lianli* and Guo, Yuyu and Zhao, Zhou and Huang, Hao and Shen, Heng Tao and Song, Jingkuan. Fine-Grained Predicates Learning for Scene Graph Generation. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022: 19445-19453, (计算视觉领域顶级会议,CCF A).
[5] Gao, Lianli and Lyu, Xinyu* and Li, Yuan-Fang and Xu, Lu and Shen, Heng Tao and Song, Jingkuan*. Informative Scene Graph Generation via Debiasing. International Journal of Computer Vision, 2025. (CCF A类期刊,影响因子:12.1,财大外文A).
[6] Zheng, Chaofan† and Lyu, Xinyu† (共同一作) and Gao, Lianli and Song, Jingkuan. Prototype-based Embedding Network for Scene Graph Generation, IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023: 22783-22792, (计算视觉领域顶级会议,CCF A).
[7] Lyu, Xinyu* and Liu, Jingwei. Local-Global Information Interaction Debiasing for Dynamic Scene Graph Generation. Artificial Intelligence and Robotics: 8th International Symposium (ISAIR), 2023, (EI).
[8] Zhao, Lei and Lyu, Xinyu and Song, Jingkuan* and Gao, Lianli. GuessWhich? Visual dialog with attentive memory network. Pattern Recognition, Volume 114, 2021: 107823–107831, (中科院一区Top期刊,影响因子:8.5,财大外文A).
[9] Gu, Yue and Lyu, Xinyu, Sun, Weijia, and Li, Weitian and Chen, Shuhong and Li, Xinyu and Ivan Marsic*. "Mutual correlation attentive factors in dyadic fusion networks for speech emotion recognition." Proceedings of the 30th ACM International Conference on Multimedia (ACM MM), 2019: 157–166, (多媒体顶级会议,CCF A).
[10] Zheng, Chaofan and Lyu, Xinyu and Guo, Yuyu and Song, Jingkuan and Gao, Lianli*. Learning to Generate Scene Graph from Head to Tail. 2022 IEEE International Conference on Multimedia and Expo (ICME), 2022: 01-06, (CCF B).
[11] Chen, Min and Lyu, Xinyu and Guo, Yuyu and Song, Jingkuan and Gao, Lianli*. Multi-Scale Graph Attention Network for Scene Graph Generation. 2022 IEEE International Conference on Multimedia and Expo (ICME), 2022: 01-06, (CCF B).
[12] Zheng, Chaofan and Gao, Lianli* and Lyu, Xinyu and Zeng, Pengpeng and Saddik Abdulmotaleb El and Shen, Heng Tao. Dual-branch hybrid learning network for unbiased scene graph generation. IEEE Transactions on Circuits and Systems for Video Technology, 2023, (中科院一区Top期刊,影响因子:8.3,财大外文A).
[13] Wang, Shuang and Gao, Lianli* and Lyu, Xinyu and Guo, Yuyu and Song, Jingkuan. Dynamic Scene Graph Generation via Temporal Prior Inference. Proceedings of the 30th ACM International Conference on Multimedia (ACM MM), 2021: 5793–5801, (多媒体顶级会议,CCF A).
[14] Zeng, Pengpeng and Gao, Lianli* and Lyu, Xinyu and Jing, Shuaiqi and Song, Jingkuan. Conceptual and Syntactical Cross-modal Alignment with Cross-level Consistency for Image-Text Matching. Proceedings of the 29th ACM International Conference on Multimedia (ACM MM), 2021:2205–2213, (多媒体顶级会议,CCF A).
[15] Zhang, Haonan and Zeng, Pengpeng and Gao, Lianli and Lyu, Xinyu and Song, Jingkuan* and Shen, Heng Tao. SPT: Spatial Pyramid Transformer for Image Captioning. IEEE Transactions on Circuits and Systems for Video Technology, 2023, (中科院一区Top期刊,影响因子:8.3,财大外文A).
其他学术成果请见个人主页:Google Scholar
主要科研项目
[1] 四川省自然科学基金青年基金项目,2025ZNSFSC1463,开放世界下的视觉关系检测算法与应用研究,2025-01至2026-12, 在研,主持
[2] 科学技术部科技创新2030-"新一代人工智能"重大项目,2018AAA0102200,视听觉信息的协同认知、主动决策与平台验证,2023-01 至 2024-06,已结题,参与
[3] 快手科技-高校科研项目,202109FKY00302,电商直播/视频的场景图生成,2020-01至2022-07,已结题,主研(排名第二)
[4] 国家自然科学基金,62020106008,协同视觉与语言处理的视觉自然认知关键技术研究,2021-01至2025-12,在研,参与
[5] 国家自然科学基金,61772116,协同深度视频理解、描述和视觉问答的关键技术研究,2018-01至2021-12,已结题,参与
[6] 国家自然科学基金,61872064,融合自然语言处理的深度视觉理解关键技术研究,2019-01至2022-12,已结题,参与
[7] J委KJ委项目-"XXX项目",61872064,信息增强的HZC目标智能认知方法",2020-09至2022-09,已结题,参与