Yuchi Liu

I am currently a final year PhD candidate at the College of Engineering, Computing and Cybernetics, Australian National University (ANU), supervised by Prof. Liang Zheng and Prof. Tom Gedeon. My research spans machine learning, computer vision, data-centric AI and generative AI. I have gained diverse industry experience through research internships at Sony AI, Baidu, Microsoft Research Asia, JD.com AI Lab and CloudWalk Technology.

Email  /  CV  /  Google Scholar  /  GitHub

• Doctor of Philosophy, Australian National University, College of Engineering, Computing and Cybernetics – Supervisor: Prof. Liang Zheng (Mar 2021 – present)
• Master of Philosophy, Australian National University, College of Engineering and Computer Science – Supervisor: Prof. Liang Zheng (Mar 2019 – Mar 2021)
• Bachelor of Software Engineering (Honours), Australian National University – Supervisor: Prof. Tom Gedeon (Mar 2017 – Dec 2018)
• Bachelor of Software Engineering, Chongqing University, College of Software Engineering (Sep 2014 – Jun 2016)

• Sony AI, PPML Group Research Intern, Tokyo – Jul 2024 – Oct 2024
  Research on agent‑based model selection.
• Baidu, Vision Technology Group Research Intern, Beijing – Jul 2022 – Jun 2023
  Model risk estimation.
• Microsoft Research Asia, Multimedia Group Research Intern, Beijing – Feb 2022 – Jul 2022
  Investigated semi‑supervised visual object segmentation.
• JD.com, AI Lab Algorithm Intern, Beijing – Aug 2019 – Mar 2020
  Worked on face recognition.
• CloudWalk Technology, Computer Vision Lab Algorithm Intern, Chongqing – Nov 2017 – Feb 2018
  Focused on model pruning for computer vision applications.

1. Towards Hierarchical Multi‑Agent Workflows for Zero‑Shot Prompt Optimization
   Yuchi Liu, Jaskirat Singh, Gaowen Liu, Ali Payani, Liang Zheng. ICLR 2025 Workshop on Reasoning and Planning for LLMs, 2025.
2. Assessing Model Generalization in Vicinity
   Yuchi Liu, Yifan Sun, Jingdong Wang, Liang Zheng. arXiv preprint, 2024.
3. Optimizing Calibration by Gaining Aware of Prediction Correctness
   Yuchi Liu, Lei Wang, Yuli Zou, James Zou, Liang Zheng. arXiv preprint, 2024.
4. A study of using synthetic data for effective association knowledge learning
   Yuchi Liu, Zhongdao Wang, Xiangxin Zhou, Liang Zheng. Multimedia Information Retrieval (20 (2): 194–206), 2023.
5. How to synthesize a large‑scale and trainable micro‑expression dataset?
   Yuchi Liu, Zhongdao Wang, Tom Gedeon, Liang Zheng. European Conference on Computer Vision (ECCV), 38–55, 2022.
6. Boosting semi‑supervised face recognition with noise robustness
   Yuchi Liu, Hailin Shi, Hang Du, Rui Zhu, Jun Wang, Liang Zheng, Tao Mei. IEEE Transactions on Circuits and Systems for Video Technology (TCSVT) 32 (2): 778–787, 2021.
7. Semi‑siamese training for shallow face learning
   Hang Du, Hailin Shi, Yuchi Liu, Jun Wang, Zhen Lei, Dan Zeng, Tao Mei. European Conference on Computer Vision (ECCV), 2020.
8. A neural micro‑expression recognizer
   Yuchi Liu, Heming Du, Liang Zheng, Tom Gedeon. IEEE Conference on Automatic Face and Gesture Recognition (FG), 1–4, 2019.
9. Generalized alignment for multimodal physiological signal learning
   Yuchi Liu, Yue Yao, Zhengjie Wang, Josephine Plested, Tom Gedeon. International Joint Conference on Neural Networks (IJCNN), 1–10, IEEE, 2019.
10. Improved techniques for building EEG feature filters
    Yue Yao, Josephine Plested, Tom Gedeon, Yuchi Liu, Zhengjie Wang. International Joint Conference on Neural Networks (IJCNN), 1–6, IEEE, 2019.

• 1st Place in the Recognition Challenge of the Second Facial Micro‑Expression Grand Challenge (MEGC2019), 2019.
• Scholarship for Outstanding Students, Chongqing University, 2015.

• Speed Up Diffusion Models (Jun 2025 – present, supervisor – Prof. Liang Zheng): Developed a training‑free method to accelerate image generation inference by multiple times for diffusion models.
• Vision‑Language Model as Detection Agency (Jul 2024 – Oct 2023, manager – Dr. Lingjuan Lyu): Post‑trained vision–language models to act as an agent that automatically invokes appropriate detectors to boost detection performance in complex environments.
• Unsupervised Model Risk Estimation (Aug 2022 – Aug 2023, manager – Dr. Yifan Sun): Proposed a framework for ranking models based on estimated risks with respect to out‑of‑distribution data.
• Visual Object Segmentation (Feb 2022 – Aug 2022, manager – Dr. Jinglu Wang): Investigated memory management challenges in semi‑supervised visual object segmentation.
• Synthetic Data for Multi‑Object Tracking (Aug 2020 – Nov 2021, supervisor – Prof. Liang Zheng): Built an engine to synthesise multi‑object tracking data with controllable factors and analysed impact factors for tracking.
• Fine‑grained Personal Re‑Identification (Sep 2020 – Oct 2021, supervisor – Prof. Liang Zheng): Extended face recognition systems to personal re‑identification tasks.
• Semi‑Supervised Face Recognition (Aug 2019 – Mar 2020, manager – Dr. Hailin Shi): Enhanced model robustness against noisy pseudo‑labels in semi‑supervised face recognition.
• Improve Micro‑Expression Recognition by Synthetic Data (Feb 2019 – Mar 2022, supervisor – Prof. Liang Zheng): Synthesised a large‑scale micro‑expression dataset and investigated computational properties of micro‑expression recognition.
• Semantic Alignment of Physiological Signals (Feb 2018 – Nov 2018, supervisor – Prof. Tom Gedeon): Aligned sub‑elements from different physiological signals into a shared feature space as part of an honours project.

• Multi‑network model training method, image annotation method and face image recognition method (2020).
• Training method and device of machine learning model, face recognition method and device (2020).