Jiawei Fan
Scholar

Jiawei Fan

Google Scholar ID: 7H674NUAAAAJ
Intel Labs China
Deep LearningComputer VisionModel CompressionDeep Learning Acceleration
Citations & Impact
All-time
Citations
26
 
H-index
2
 
i10-index
1
 
Publications
6
 
Co-authors
5
list available
Resume (English only)
Academic Achievements
  • ICML 2025: Proposed Morse, a universal diffusion acceleration framework applicable to any diffusion model.
  • NeurIPS 2024 (first author): Introduced ScaleKD, the first work to transfer scalability of pre-trained ViTs to diverse student architectures.
  • NeurIPS 2023 (first author): Proposed Af-DCD, an augmentation-free dense contrastive distillation framework for efficient semantic segmentation.
  • ACCV 2022 (co-first author): Developed TCVM for self-supervised video representation learning; received Best Paper Award Honorable Mention (Oral).
  • ACM MM 2022 (first author): Proposed DTR, a parameter-free information bottleneck regularization framework for action recognition.
  • ICPR 2022 (first author): Proposed Episodic Projection Network for out-of-distribution detection in few-shot learning.
  • Ranked 7th in the 3rd Person in Context Workshop at CVPR 2021.
  • Recipient of Intel Labs China DRA Award (2023), MEGVII Outstanding Intern Award (2022), and BUPT First-Class Scholarship (2021).