Guosheng Hu is currently a Senior Lecturer of University of Bristol (Sep 2024 - ). Before that, He was the Head of Research of Oosto (formerly Anyvision), a leading visual AI company (2016 - 2024). He leads a wide range of research on computer vision and machine learning, and many of his research endeavors have successfully been commercialsed.
Professional Experience
Senior Lecturer: Guosheng Hu joined University of Bristol as a Senior Lecturer of AI in September, 2024 [link].
Industrial Roles: In 2016, Guosheng Hu joined Oosto (formerly Anyvision) as a senior researcher intially and then was promoted as Head of Research. His innovative research have resulted in numerous publications and commercially successful products at Oosto.
Postdoc: Following his PhD, Guosheng Hu conducted postdoctoral research under the supervision of Dr. Jakob Verbeek in the LEAR group (now THOTH group) at INRIA, Grenoble from 2015 to 2016.
PhD: Guosheng Hu earned his PhD under the supervision of Prof. Josef Kittler at the University of Surrey from 2010 to 2015. His doctoral research focused on advancing 2D and 3D face modeling techniques, leading to an internship at Disney Research, Pittsburgh, where he worked alongside Prof. Yisong Yue and Dr. Jose Rafael Tena.
I (try to) adhere to the principles of Slow Science.
News
01/07/24 One ECCV 2024 paper is accepted.
27/02/24 Two CVPR 2024 papers are accepted and one is ‘highlight’.
22/02/24 I serve as an Associate Editor of Neurocomputing.
09/12/23 Three AAAI 2024 papers are accepted.
Recruitment
I am looking for prospective PhD students, visiting students, and interns for the 2025 entry. Research topics focus on computer vision and model acceleration. Scholarships (University-level, Faculty-level, CDTs, CSC and others) are available for both home and international students. Feel free to email me to discuss these opportunities!
Research
Mission: Guosheng Hu’s research aims to reduce AI’s carbon footprint through model acceleration techniques.
Commercial Success: Guosheng Hu has led the development of several successful products that accelerate various computer vision models on diverse hardware platforms, including Nvidia’s A100, V100, Jetson Orin, Intel CPUs, and Ambaralla Chips.
Publications: His work is widely published in prestigious journals and conferences [Google Scholar], reflecting his broad and in-depth knowledge of computer vision and machine learning tasks. Recently, his research interests have expanded to the acceleration of training and inference for foundation models, such as large language models and large multimodal models.
Selected Publications
Model Acceleration
Neural Architecture Search and AutoML
Li, Yonggang, Guosheng Hu, Yongtao Wang, Timothy Hospedales, Neil M. Robertson, and Yongxin Yang. “Differentiable automatic data augmentation.” ECCV, 2020. [code]
Liang, Tingting, Yongtao Wang, Zhi Tang, Guosheng Hu, and Haibin Ling. “Opanas: One-shot path aggregation network architecture search for object detection.” CVPR, 2021. [code]
Fan, Zhenkun, Guosheng Hu, Xin Sun, Gaige Wang, Junyu Dong, and Chi Su. “Self-attention neural architecture search for semantic image segmentation.” Knowledge-Based Systems 239 (2022): 107968.
Knowledge Distillation, Quantisation and Pruning
Jin, Yufeng, Guosheng Hu, Haonan Chen, Duoqian Miao, Liang Hu, and Cairong Zhao. “Cross-modal distillation for speaker recognition.” AAAI, 2023.
Sun, Tianli, Haonan Chen, Guosheng Hu, and Cairong Zhao. “Explainability-Based Knowledge Distillation.” Available at SSRN 4460609.
Parameter-Efficient Tuning
Chen, Hao, Ran Tao, Han Zhang, Yidong Wang, Wei Ye, Jindong Wang, Guosheng Hu, and Marios Savvides. “Conv-adapter: Exploring parameter efficient transfer learning for convnets.” CVPR Workshop Prompting in Vision, 2024. [code coming soon]
Computer Vision
Detection, Classification and Segmentation
Sun, Guanxiong, Yang Hua, Guosheng Hu, and Neil Robertson. “Efficient one-stage video object detection by exploiting temporal consistency.” ECCV, 2022. [slides] [code]
Sun, Guanxiong, Yang Hua, Guosheng Hu, and Neil Robertson. “Tdvit: Temporal dilated video transformer for dense video tasks.” In European Conference on Computer Vision, pp. 285-301. Cham: Springer Nature Switzerland, 2022. [slides] [code]
Sun, Guanxiong, Yang Hua, Guosheng Hu, and Neil Robertson. “Mamba: Multi-level aggregation via memory bank for video object detection.” AAAI. 2021. [slides] [code]
Liang, Tingting, Yongtao Wang, Zhi Tang, Guosheng Hu, and Haibin Ling. “Opanas: One-shot path aggregation network architecture search for object detection.” CVPR, 2021. [code]
Chen, Zhiyang, Yousong Zhu, Chaoyang Zhao, Guosheng Hu, Wei Zeng, Jinqiao Wang, and Ming Tang. “Dpt: Deformable patch-based transformer for visual recognition.” In Proceedings of the 29th ACM International Conference on Multimedia, pp. 2899-2907. 2021. [code]
Lu, Bingxu, Qinghua Hu, Yu Wang, and Guosheng Hu. “Rcanet: Row-column attention network for semantic segmentation.” In ICASSP 2022-2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 2604-2608. IEEE, 2022.
Fan, Zhenkun, Guosheng Hu, Xin Sun, Gaige Wang, Junyu Dong, and Chi Su. “Self-attention neural architecture search for semantic image segmentation.” Knowledge-Based Systems 239 (2022): 107968.
Face Analysis
Hu, Guosheng, Yongxin Yang, Dong Yi, Josef Kittler, William Christmas, Stan Z. Li, and Timothy Hospedales. “When face recognition meets with deep learning: an evaluation of convolutional neural networks for face recognition.” ICCV workshops, 2015.
Hu, Guosheng, Xiaojiang Peng, Yongxin Yang, Timothy M. Hospedales, and Jakob Verbeek. “Frankenstein: Learning deep face representations using small data.” IEEE Transactions on Image Processing 27, no. 1 (2017): 293-303.
Hu, Guosheng, Yang Hua, Yang Yuan, Zhihong Zhang, Zheng Lu, Sankha S. Mukherjee, Timothy M. Hospedales, Neil M. Robertson, and Yongxin Yang. “Attribute-enhanced face recognition with neural tensor fusion networks.” ICCV, 2017.
Mu, Guodong, Di Huang, Guosheng Hu, Jia Sun, and Yunhong Wang. “Led3d: A lightweight and efficient deep approach to recognizing low-quality 3d faces.” CVPR, 2019.
Hu, Guosheng, Fei Yan, Chi-Ho Chan, Weihong Deng, William Christmas, Josef Kittler, and Neil M. Robertson. “Face recognition using a unified 3D morphable model.” ECCV, 2016.
Liu, Zhiwei, Xiangyu Zhu, Guosheng Hu, Haiyun Guo, Ming Tang, Zhen Lei, Neil M. Robertson, and Jinqiao Wang. “Semantic alignment: Finding semantically consistent ground-truth for facial landmark detection.” CVPR, 2019.
Chen, Haonan, Guosheng Hu, Zhen Lei, Yaowu Chen, Neil M. Robertson, and Stan Z. Li. “Attention-based two-stream convolutional networks for face spoofing detection.” IEEE Transactions on Information Forensics and Security 15 (2019): 578-593.
Zhao, Cairong, Chutian Wang, Guosheng Hu, Haonan Chen, Chun Liu, and Jinhui Tang. “ISTVT: interpretable spatial-temporal video transformer for deepfake detection.” IEEE Transactions on Information Forensics and Security 18 (2023): 1335-1348.
Wen, Xin, Biying Li, Haiyun Guo, Zhiwei Liu, Guosheng Hu, Ming Tang, and Jinqiao Wang. “Adaptive variance based label distribution learning for facial age estimation.” ECCV, 2020.
Hu, Guosheng, Li Liu, Yang Yuan, Zehao Yu, Yang Hua, Zhihong Zhang, Fumin Shen et al. “Deep multi-task learning to recognise subtle facial expressions of mental states.” ECCV, 2018.
Foundation Model Applications
Wang, Tianfu, Guosheng Hu, and Hongguang Wang. “Object Pose Estimation via the Aggregation of Diffusion Features.” CVPR, 2024. [code]
Li, Wen, Yuyang Yang, Shangshu Yu, Guosheng Hu, Chenglu Wen, Ming Cheng, Cheng Wang, “DiffLoc: Diffusion Model for Outdoor LiDAR Localization.” CVPR, 2024.
Song, Zifan, Guosheng Hu, and Cairong Zhao. “Diverse Person: Customize Your Own Dataset for Text-Based Person Search.” AAAI, 2024.
Machine Learning
Multimodal Learning
Jin, Yufeng, Guosheng Hu, Haonan Chen, Duoqian Miao, Liang Hu, and Cairong Zhao. “Cross-modal distillation for speaker recognition.” AAAI, 2023.
Wang, Hao, Shengda Luo, Guosheng Hu, and Jianguo Zhang. “Gradient-Guided Modality Decoupling for Missing-Modality Robustness.” AAAI, 2024.
Metric Learning
Wang, Xinshao, Yang Hua, Elyor Kodirov, Guosheng Hu, Romain Garnier, and Neil M. Robertson. “Ranked list loss for deep metric learning.” CVPR, 2019.
Wang, Xinshao, Yang Hua, Elyor Kodirov, Guosheng Hu, and Neil M. Robertson. “Deep metric learning by online soft mining and class-aware attention.” AAAI. 2019.
Jiang, Xiruo, Sheng Liu, Xili Dai, Guosheng Hu, Xingguo Huang, Yazhou Yao, Guo-Sen Xie, and Ling Shao. “Deep metric learning based on meta-mining strategy with semiglobal information.” IEEE Transactions on Neural Networks and Learning Systems (2022).
Noisy Label
Wang, Zhen, Guosheng Hu, and Qinghua Hu. “Training noise-robust deep neural networks via meta-learning.” CVPR, 2020.
Sun, Zeren, Xian-Sheng Hua, Yazhou Yao, Xiu-Shen Wei, Guosheng Hu, and Jian Zhang. “Crssc: salvage reusable samples from noisy data for robust learning.” In Proceedings of the 28th ACM international conference on multimedia, pp. 92-101. 2020.
Domain Adaptation
Gao, Jian, Yang Hua, Guosheng Hu, Chi Wang, and Neil M. Robertson. “Discrepancy-guided domain-adaptive data augmentation.” IEEE Transactions on Neural Networks and Learning Systems 34, no. 8 (2021): 5064-5075.
Gao, Jian, Yang Hua, Guosheng Hu, Chi Wang, and Neil M. Robertson. “Reducing distributional uncertainty by mutual information maximisation and transferable feature learning.” ECCV, 2020.
Others
Song, Zifan, Xiao Gong, Guosheng Hu, and Cairong Zhao. “Deep perturbation learning: enhancing the network performance via image perturbations.” ICML, 2023.
Wang, Tianyang, Xingjian Li, Pengkun Yang, Guosheng Hu, Xiangrui Zeng, Siyu Huang, Cheng-Zhong Xu, and Min Xu. “Boosting active learning via improving test performance.” AAAI, 2022.
Mai, Zhijun, Guosheng Hu, Dexiong Chen, Fumin Shen, and Heng Tao Shen. “Metamixup: Learning adaptive interpolation policy of mixup with metalearning.” IEEE transactions on neural networks and learning systems 33, no. 7 (2021): 3050-3064.
Gong, Xiao, Guosheng Hu, Timothy Hospedales, and Yongxin Yang. “Adversarial robustness of open-set recognition: face recognition and person re-identification.” ECCV Workshop, 2020.
Hu, Guosheng, Li Liu, Yang Yuan, Zehao Yu, Yang Hua, Zhihong Zhang, Fumin Shen et al. “Deep multi-task learning to recognise subtle facial expressions of mental states.” ECCV, 2018.
Hu, Guosheng, Yuxin Hu, Kai Yang, Zehao Yu, Flood Sung, Zhihong Zhang, Fei Xie et al. “Deep stock representation learning: From candlestick charts to investment decisions.” In 2018 IEEE international conference on acoustics, speech and signal processing (ICASSP), pp. 2706-2710. IEEE, 2018.
Sun, Tianli, Haonan Chen, Guosheng Hu, Lianghua He, and Cairong Zhao. “Explainability of Speech Recognition Transformers via Gradient-based Attention Visualization.” IEEE Transactions on Multimedia (2023).
Supervision
“Education is not the filling of a pail, but the lighting of a fire.” - William Butler Yeats
I work as an industrial supervisor to co-supervise Oosto sponsored PhD students:
Hao Chen, Carnegie Mellon University (2022 - )
Uzair Ahmed, Carnegie Mellon University (2022 -)
Ran Tao, Carnegie Mellon University (2022 -)
Guanxiong Sun, Queen’s University Belfast (2019 - 2022)
Jian Gao, Queen’s University Belfast (2018 - 2022)
Xinshao Wang, Queen’s University Belfast (2017 - 2020)
Funding and Grants
King’s College London NMES Enterprise Engagement Partnerships Fund, January - June 2024.
Fellowship for entrepreneurs of University of Surrey, August 2024 - July 2025.
Professional Engagements
IEEE Senior Member (2022 - )
Associate Editor, Neurocomputing (2024 -)
Associate Editor, IET Image Processing (2021 - )
Area Chair: BMVC 2022, 2023
Programme Committee: CVPR, ICCV, ECCV, NIPS, ICML, AAAI, ICLR
Tutorial and Workshop: FG 2018, 2019, 2023
Conclusion
Guosheng Hu’s impact on the fields of computer vision and machine learning is profound. As the Head of Research at Oosto and an Honorary Professor of Practice at Queen’s University Belfast, his work continues to shape the future of visual AI. Through his innovative research, dedicated mentorship, and professional contributions, Hu exemplifies the transformative potential of AI in addressing complex challenges and driving technological progress.
Contact
- huguosheng100 at gmail.com
- 48-60 High St, Belfast, BT1 2BE, United Kingdom
