胡庆浩      中国科学院自动化研究所 副研究员


现为中国科学院自动化研究所 副研究员、硕导，研究领域主要包括深度模型轻量化、具身智能等方向，发表IEEE T-NNLS、IEEE T-MM等权威期刊和ICLR、CVPR、ICCV、AAAI等顶级会议论文三十余篇。主持国家自然科学基金青年基金、北京市自然基金-昌平联合创新基金、科技部2030新一代人工智能重大项目子课题、国家电网企业横向委托项目等科研项目，参与中科院战略先导、国家重点研发计划等多研发项目，现为中国图象图形学会会员、CSIG视觉大数据专委会委员。曾获2015年MSR-Bing图像识别挑战赛冠军、2019年英伟达奖学金、2019年ICCV轻量化人脸识别挑战赛亚军、人脸实时检测挑战赛亚军，入选2024年CCF-百度松果基金学者，获得2024年电子工业出版社年度优秀作者奖、2025年第三届“智天论坛”优秀论文、2025年电子学会科技进步一等奖。


电子邮件： huqinghao2014@ia.ac.cn

通信地址： 北京市海淀区中关村东路95号

邮政编码： 100190

研究领域

  • 深度神经网络轻量化计算：研究神经网络的高效训练与推理算法、大模型参数高效微调算法、大模型量化剪枝方法。
  • 具身智能：研究具身智能高效计算方法、具身大模型架构设计、具身智能在线进化算法。

招生信息

招生专业：

081104-模式识别与智能系统

081203-计算机应用技术


招生方向：

大模型加速与压缩、具身智能、大模型高效训练与微调方法


实习生招聘：

长期欢迎优秀本科生、研究生线上、线下实习

教育背景

2014-09--2019-07   中国科学院自动化研究所   博士
2010-09--2014-07   西北工业大学   学士

工作经历

2022-04~现在, 中国科学院自动化研究所, 副研究员

2019-07~2022-04,中国科学院自动化研究所, 助理研究员

出版信息

   
图书著作
  • 《深度神经网络高效计算：大模型轻量化原理与关键技术》, 电子工业出版社, 2024-08, 王培松、胡庆浩、莫子韬编著，程健主编。
发表论文
  1. YuanhuiWang, Kunlong Liu1, Minnan Pei, Zhangming Li, Peisong Wang, Qinghao Hu↑.MemeBQ: Memory Efficient Binary Quantization of LLMs [C]//The 39th AAAI Conference on Artificial Intelligence（AAAI）, 2026,通讯作者. [code]
  2. Changwang Mei, Peisong Wang, Shuang Qiu, Gang Li, Qinghao Hu, Yifan Zhang, Zhihui Wei, Jian Cheng.HiCache: hierarchical timestep-aware caching for diffusion transformer acceleration. Visual Intelligence, 2026
  3. Jiahe Qian, Peisong Wang, Zhengyang Zhuge, Qinghao Hu, Jian Cheng. A Universal Self-Attention Enhancement for Bridging Low-bit Quantization and Vision Transformers. WACV 2026
  4. Piehuan Ni, Zitao Mo, Tielong Liu, Hongli Wen, Zeyu Zhu, Minnan Pei, Junwen Si, Weifan Guan, Peisong Wang, Qinghao Hu, Gang Li, Jian Cheng. APEX: Integer-only Non-linear Function Approximation for Efficient Cross-Modal Inference. DATE 2026.
  5. Yiqun Chen, Qiang Chen, Qinghao Hu, Jian Cheng. Dual Assignment of Labels for End-to-end Fully Convolutional Object Detection.Pattern Recognition, Vol.129, 2026,[code]
  6. Weifan Guan, Qinghao Hu, Aosheng Li, Jian Cheng.Efficient vision-language-action models for embodied manipulation: A systematic survey. Preprint. 首篇高效VLA综述, [code]
  7. Xingting Yao*, Qinghao Hu*, Fei Zhou, Tielong Liu, Gang Li, Peisong Wang, Jian Cheng. Towards Efficient and Accurate Spiking Neural Networks via Adaptive Bit Allocation. Neural Networks, 2025, 中科院JCR一区,共同一作
  8. Jiayao Ling*, Gang Li, Qinghao Hu*, Xiaolong Lin, Cheng Gu, Jian Cheng, Xiaoyao Liang. SBQ: Exploiting Significant Bits for Efficient and Accurate Post-Training DNN Quantization. Design, Automation & Test in Europe Conference (DATE) 2025, 共同一作.
  9. Xingting yao*, Qinghao Hu*, Fei Zhou, Tielong Liu, Zitao Mo, Zeyu Zhu, Zhengyang Zhuge, Jian Cheng. SpiNeRF: Direct-trained Spiking Neural Networks for Efficient Neural Radiance Field Rendering. Frontiers in Neuroscience, July 23, 2025,共同一作.[code]
  10. Zhangming Li, Qinghao Hu↑, Yiqun Chen, Peisong Wang, Yifan Zhang, Jian Cheng↑. LoRaDA: Low-Rank Direct Attention Adaptation for Efficient LLM Fine-tuning. EMNLP 2025 Findings, 共同通讯
  11. Zeyu Zhu, Peisong Wang, Qinghao Hu, Gang Li, Xiaoyao Liang, Jian Cheng. FastGL: A GPU-Efficient Framework for Accelerating Sampling-Based GNN Training at Large Scale.ASPOS 2024
  12. Xing Lan, Jiayi Lyu, Kun Dong, Hanyu Jiang, Qinghao Hu, Jian Xue. Does pixel value represent facial landmark well in heatmap?IEEE Transactions on Circuits and Systems for Video Technology(TCSVT), vol. 34, no. 12, pp. 13016-13028, Dec. 2024
  13. Zeyu Zhu, Fanrong Li, Gang Li, Zejian Liu, Zitao Mo, Qinghao Hu, Xiaoyao Liang, Jian Cheng.Mega: A memory-efficient gnn accelerator exploiting degree-aware mixed-precision quantization.2024 IEEE International Symposium on High-Performance Computer Architecture (HPCA).
  14. Zhixiang Ye*,Qinghao Hu* Tianli Zhao, Wangping Zhou,Jian Cheng. MCUNeRF: Packing NeRF into an MCU with 1MB Memory. ACM International Conference on Multimedia(ACM MM) ,2023,共同一作.
  15. Tianli Zhao, Qinghao Hu, Xiangyu He, Weixiang Xu, Jiaxing Wang, Cong Leng, Jian Cheng. ECBC: Efficient Convolution via Blocked Columnizing. IEEE Transactions on Neural Networks and Learning Systems (TNNLS), Vol.34, No.1, pp.433-445, 2023.,中科院JCR一区
  16. Xing Lan, Qinghao Hu, Jian Cheng. ATF: An Alternating Training Framework for Weakly Supervised Face Alignment. IEEE Transactions on Multimedia, Vol.25, pp.1798-1809, 2023.中科院JCR一区
  17. Xiangyu Chen, Qinghao Hu, Kaidong Li, Cuncong Zhong, Guanghui Wang. Accumulated Trivial Attention Matters in Vision Transformers on Small Datasets. Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2023, pp. 3984-3992
  18. Zeyu Zhu, Fanrong Li, Zitao Mo, Qinghao Hu, Gang Li, Zejian Liu, Xiaoyao Liang, Jian Cheng. A2Q: Aggregation-Aware Quantization for Graph Neural Networks. ICLR 2023.
  19. Qiang Chen,Qiman Wu, Jian Wang, Qinghao Hu↑, Tao Hu, Errui Ding, Jian Cheng, Jingdong Wang. Mixformer: Mixing features across windows and dimension. IEEE Conference on Computer Vision and Pattern Recognition (CVPR),2022,通讯作者.[code]
  20. Qinghao Hu, Gang Li, Qiman Wu, Jian Cheng.PalQuant: Accelerating High-Precision Networks on Low-Precision Accelerator.European Conference on Computer Vision (ECCV), 2022,一作.[code]
  21. Weixiang Xu, Xiangyu He, Tianli Zhao, Qinghao Hu, Peisong Wang, Jian Cheng. Soft Threshold Ternary Networks. IJCAI 2020.
  22. Xing Lan, Qinghao Hu, Fangzhou Xiong, Cong Leng, Jian Cheng. ATF: Towards Robust Face Alignment via Leveraging Similarity and Diversity across Different Datasets. ACM MM 2020.
  23. Xiangyu He, Zitao Mo, Ke Cheng, Weixiang Xu, Qinghao Hu, Peisong Wang, Qingshan Liu, Jian Cheng. Proxybnn: Learning binarized neural networks via proxy matrices. ECCV 2020.
  24. Fanrong Li, Zitao Mo, Peisong Wang, Zejian Liu, Jiayun Zhang, Gang Li, Qinghao Hu, Xiangyu He, Cong Leng, Yang Zhang, Jian Cheng. A System-Level Solution for Low-Power Object Detection. ICCV 2019 Workshop on Low-Power Computer Vision.
  25. Xianyang Li, Feng Wang, Qinghao Hu, Cong Leng. Airface: Lightweight and efficient model for face recognition. ICCV 2019 Workshop . (Ranked the 2nd place on Lightweight Face Recognition Challenge )
  26. Jian Cheng, Peisong Wang, Gang Li, Qinghao Hu, Hanqing Lu. Recent Advances in Efficient Computation of Deep Convolutional Neural Networks. Frontiers of Information Technology & Electronic Engineering (FITEE), Vol.19, No.1, pp.64-77, 2018.
  27. Peisong Wang, Qinghao Hu, Zhiwei Fang, Chaoyang Zhao, Jian Cheng. DeepSearch: A Fast Image Search Framework for Mobile Devices. ACM Transactions on Multimedia Computing Communications and Applications (TOMM), Vol.14(1), 2018.
  28. Peisong Wang, Qinghao Hu, Yifang Zhang, Chunjie Zhang, Yang Liu and Jian Cheng. Two-Step Quantization for Low-bit Neural Networks. CVPR 2018.
  29. Qinghao Hu,Gang Li, Peisong Wang, Yifan Zhang, Jian Cheng. Training Binary Weight Networks via Semi-Binary Decomposition. European Conference on Computer Vision (ECCV), 2018, 一作
  30. Qinghao Hu,Peisong Wang, Jian Cheng. From Hashing to CNNs: Training Binary Weight Networks via Hashing. The Thirty-Second AAAI Conference on Artificial Intelligence (AAAI), 2018, 一作
  31. Peisong Wang, Qinghao Hu, Zhiwei Fang, Chaoyang Zhao and Jian Cheng. DeepSearch: A fast image search framework for mobile devices. ACM Transactions on Multimedia Computing, Communications, and Applications (TOMM), 2018.
  32. Guan’An Wang, Qinghao Hu,Yang Yang,Jian Cheng, Zeng-Guang Hou. Adversarial Binary Mutual Learning for Semi-Supervised Deep Hashing.IEEE Transactions on Neural Networks and Learning Systems (TNNLS),2021,二作,中科院JCR一区
  33. Jian Cheng, Jiaxiang Wu, Cong Leng, Yuhang Wang, Qinghao Hu. Quantized CNN: A Unified Approach to Accelerate and Compress Convolutional Networks. IEEE Transactions on Neural Networks and Learning Systems (TNNLS), Vol.29, No.10, pp.4730-4743, 2018.
  34. Qinghao Hu, Jiaxiang Wu, Lu Bai, Yifan Zhang, Jian Cheng. Fast K-Means for Large Scale Clustering. ACM CIKM 2017.
  35. Qinghao Hu, Jiaxiang Wu, Jian Cheng, Lifang Wu, Hanqing Lu. Pesudo Label based Unsupervised Deep Discriminative Hashing for Image Retrieval. ACM MM 2017.
  36. Jiaxiang Wu, Cong Leng, Yuhang Wang, Qinghao Hu, Jian Cheng. Quantized Convolutional Neural Networks for Mobile Devices. CVPR 2016. [code]. 1700+ google citations.
  37. Jiaxiang Wu, Qinghao Hu, Cong Leng, Jian Cheng. Shoot to Know What: An Application of Deep Networks on Mobile Devices. AAAI 2016 (Demo track).
  38. Qiang Song, Sixie Yu, Cong Leng, Jiaxiang Wu, Qinghao Hu, Jian Cheng. Learning Deep Features for MSR-Bing Information Retrieval Challenge. ACM Multimedia 2015. (Ranked the 1st place on Visual Recognition task and 3rd place on Image Retrieval task)

科研活动

主要承担科研项目：

  1. 面向MCU的二值多模态神经网络计算关键技术研究，北京市自然科学基金-昌平创新联合基金，负责人，2024/10-2027/09
  2. 面向移动计算的深度神经网络自动轻量化研究，国家自然科学基金青年项目，负责人，2022/01-2024/12
  3. 博弈智能对抗演练场，科技部科技创新2030重大项目，子课题负责人，2022/12-2025/11
  4. 国产自主可控多模态大模型关键技术，北京市科技计划项目,任务负责人，2023/09- 2025/09
  5. 基于二值量化的大模型高效推理技术研究，CCF-百度松果基金，负责人，2024/12-2025/11
  6. 输变电设备可视缺陷识别模型前端化移植技术，国家电网横向委托，负责人，2022.1-2023.12
  7. 电力异构融合类脑计算关键技术研究，国家电网科技项目，课题负责人，2023.10-2025.12
  8. 基于图文模型的多目标可解释现场作业违章行为识别关键技术研究及应用，国家电网科技项目，课题负责人，2024.01-2025.12