Linwei Tao is a Ph.D student at The University of Sydney, supervised by A/Prof Chang Xu. He received B.Eng degree from Huazhong University of Science and Technology in 2017. His research focuses on Confidence Calibration and Uncertainty Estimation in Deep Learning and Large (Vision) Language Models. He also worked as a head TA for COMP5329 Deep Learning and HTIN5005 Applied Healthcare Data Science in USYD.

Google Scholar / Github / CV

🔥 News

• 2025.05: 🎉🎉 One paper is accepted by ICML 2025.
• 2025.02: 🎉🎉 One paper is accepted by CVPR 2025.
• 2025.01: 🎉🎉 One paper is accepted by ICLR 2025.
• 2024.11: 🎉🎉 One paper is accepted by AAAI 2025.

📝 Publications

• Beyond One-Hot Labels: Semantic Mixing for Model Calibration
  Haoyang Luo, Linwei Tao, Minjing Dong, Chang Xu.
  ICML, 2025.

• Uncertainty Weighted Gradients for Model Calibration
  Jinxu Lin, Linwei Tao, Minjing Dong, Chang Xu.
  CVPR, 2025.

• KG-UQ: Knowledge Graph-Based Uncertainty Quantification for Long Text in Large Language Models
  Yingqing Yuan, Linwei Tao, Haohui Lu, Matloob Khushi, Imran Razzak, Mark Dras,Jian Yang, Usman Naseem.
  WWW Workshop, 2025.

• Diffusion Attribution Score: Evaluating Training Data Influence in Diffusion Model
  Jinxu Lin, Linwei Tao, Minjing Dong, Chang Xu.
  ICLR, 2025. (Spotlight, 5.1% of 11,500 submissions) [Paper]

• Feature Clipping for Uncertainty Calibration
  Linwei Tao, Minjing Dong, Chang Xu.
  AAAI, 2025. [Paper] [Code]

• GraphFusion: Integrating Multi-Level Semantic Information with Graph Computing forEnhanced 3D Instance Segmentation
  Lei Pan, Wuyang Luan, Yuan Zheng, Junhui Li, Linwei Tao, Chang Xu.
  Neurocomputing, 2024. [Paper] [Code]

• A Benchmark Study on Calibration
  Linwei Tao, Younan Zhu, Haolan Guo, Minjing Dong, Chang Xu.
  ICLR, 2024. [Paper] [Dataset] [Code] [Project Page] [Poster] [Slides]

• Dual Focal Loss for Calibration
  Linwei Tao, Minjing Dong, Chang Xu.
  ICML, 2023. [Paper] [Code] [Poster] [Slides]

• Calibrating a Deep Neural Network with Its Predecessors
  Linwei Tao, Minjing Dong, Daochang Liu, Changming Sun, Chang Xu.
  IJCAI, 2023. [Paper] [Code] [Playground] [Slides]

• Mitigating Object Hallucinations in Large Vision-Language Models via Attention Calibration
  Younan Zhu, Linwei Tao, Minjing Dong, Chang Xu.
  Preprint, 2025. [Paper]

• Consistency Calibration: Improving Uncertainty Calibration via Consistency among Perturbed Neighbors
  Linwei Tao, Haolan Guo, Minjing Dong, Chang Xu.
  Preprint, 2025. [Paper]

• Visual Imitation Learning with Calibrated Contrastive Representation
  Yunke Wang, Linwei Tao, Bo Du, Yutian Lin, Chang Xu.
  Preprint, 2024. [Paper]

📖 Education

• 2023.10 - Present, Ph.D in Computer Vision, University of Sydney
• 2021.10 - 2023.03, M.Phil in Computer Vision, University of Sydney
• 2020.02 - 2021.07, Master of Data Science, University of Sydney
• 2013.09 - 2017.06, B.Eng in Communication and Engineering, Huazhong University of Science and Technology

🧑‍🏫 Teaching

• 2024 S2, Head TA of OCMP5329 Deep Learning (Online)
• 2022 S1, 2023 S1, 2024 S1, Head TA of COMP5329 Deep Learning
• 2022 S2, 2023 S2, Head TA of HTIN5005 Applied Healthcare Data Science

💼 Internship

• 2015.06 - 2015.09, IOS developer, Ctrip.com
• 2014.06 - 2014.09, Frontend developer, Sunallies.com

🎖 Honors and Awards

• 2024.05, Google Cloud Research Credits Award ($2340 AUD)
• 2023.06, International Tuition Fee Scholarship, University of Sydney
• 2023.05, Faculty of Engineering Research Support Scholarship, University of Sydney
• 2015.09, Excellent Student Cadre, Huazhong University of Science and Technology
• 2014.06, Science and Technology Scholarship, Huazhong University of Science and Technology

💬 Invited Talks

• 2024.10, “ITCD: Image to Text Translation for Classification by Diffusion Models”, at Efficiency, Security, and Generalization of Multimedia Foundation Models Workshop @ ACM Multimedia 2024 [Link]
• 2023.05, “Dual Focal Loss for Calibration”, at AI-Time [Slides]
• 2023.02, “Calibrating a Deep Neural Network with Its Predecessors”, at University of Sydney, Deep Learning class [Slides]

💻 Academic Service

• Conference Reviewer: ICML, NeuIPS, AAAI, ICLR, CVPR, ICML, IJCAI, ACMMM, AISTATS
• Journal Reviewer: T-MM, TMLR, DAMI, TPAMI