CV

You can download my resume here.

Education

• Duke University, Durham, North Carolina, USA, 2023 -
  Ph.D. Candidate, Vision and Image Processing Lab, Department of Biomedical Engineering
  • GPA: 3.84/4.0
• Shanghai Jiao Tong University (SJTU), Shanghai, China, 2019 - 2023
  Bachelor’s Degree of Engineering, major in Biomedical Engineering, minor in Computer Science and Engineering
  • GPA: 3.84/4.3 (Top 10%)
  • See more on my final transcript

Scholar experience

• Aug. 2023 – Present: VIP Lab @ Duke directed by Sina Farsiu

  PRIMA: Pre-training with Risk-integrated Image-Metadata Alignment for Medical Diagnosis via LLM

  • Elevated clinical metadata into structured semantic knowledge by fine-tuning ClinicalBERT on RAG-constructed medical corpora, explicitly incorporating domain priors without reliance on large-scale paired image–text datasets.
  • Proposed a unified multimodal pre-training framework with four complementary loss functions to coordinate global–local feature alignment across heterogeneous clinical modalities.
  • Developed an end-to-end pipeline leveraging Qwen3 for multimodal feature synthesis and alignment, achieving state-of-the-art performance and strong cross-dataset generalization on PAD-UFES-20 and AQUA.
  • Abstract accepted as an Oral Presentation at ARVO 2026; full paper submitted to MICCAI 2026 and available on arXiv; extended journal manuscript in preparation.

  A multi-granularity language learning approach to boost visual understanding

  • Proposed a novel contrastive learning framework enabling joint multi-label and cross-granularity alignment for medical image understanding.
  • Designed multi-granular learning objectives to improve representation consistency across label hierarchies.
  • Constructed large-scale multi-granular retinal and X-ray image-text datasets for systematic evaluation.
  • Accepted by ICLR 2026; available on arXiv.

  An Automated Quantitative Ulcer Analysis (AQUA) algorithm to classify Microbial keratitis (MK) organism types

  • Proposed a contrastive-learning-based approach to extract robust representations across heterogeneous clinical data patterns.
  • Developed a triple-stage multimodal framework to integrate imaging, metadata, and clinical features for organism-type classification.
  • Journal manuscript in preparation (expected 2026).

• June. 2022 – Nov. 2022: CCVL @ JHU directed by Alan Yuille & VLAA @ UCSC directed by Yuyin Zhou & Cihang Xie

  Multi-view MAE for 3D medical image representation learning

  • Proposed the first multi-view framework for self-supervised 3D medical image representation learning.
  • Achieved performance comparable to state-of-the-art methods with substantially reduced training cost.
  • Published in MICCAI 2023 (Oral Presentation).

• Feb. 2022 – Jan. 2023: Advanced MRI Lab @ SJTU directed by Hongjiang Wei

  Brain Region Segmentation and Age Estimation Using QSM

  • Developed a novel segmentation network for key brain regions on QSM images to support brain age estimation.
  • Improved brain age prediction accuracy compared to prior T1-weighted MRI-based approaches.
  • Published in ISMRM 2023 and IEEE Journal of Biomedical and Health Informatics (JBHI).

Publications

2026

Y Wang, C He, MC Lu, M Pawar, L Niziol, M Woodward, S Farsiu. PRIMA: Pre-training with Risk-integrated Image-Metadata Alignment for Medical Diagnosis via LLM. ArXiv:2602.23297.

Z Li¹, Y Wang¹, S Farsiu, P Kinahan. Boosting Medical Visual Understanding From Multi-Granular Language Learning. The Fourteenth International Conference on Learning Representations (ICLR).

Y Wang, C He, LM Niziol, M Pawar, M Lu, MA Woodward, & S Farsiu. Large-Language-Model-Enhanced Multi-Modal Subtype Diagnosis of Microbial Keratitis on Slit-Lamp Photographs and Metadata. Association for Research in Vision and Ophthalmology (ARVO) 2026 Annual Meeting.

J Ong, MC Lu, C Thanitcul, M Pawar, JN Hart, EL Vogt, C Deng, S Farsiu, Y Wang, P Dmitriev, A Gupta. Deep Learning-Based Classification of Slit-Lamp Photograph Quality in Microbial Keratitis. Ophthalmology Science. 2026 Jan 21:101086.

2025

Z Li¹, Y Wang¹, S Farsiu, P Kinahan. Large‑scale data harmonization of radiology studies via multigranular vision‑language alignment. SPIE Medical Imaging 2026.

J Ong, M Lu, C Thanitcul, M Pawar, JN Hart, E Vogt, S Farsiu, Y Wang, P Dmitriev, A Gupta, N Nallasamy & MA Woodward. Automated Deep Learning Classification of the Quality of Slit-Lamp Photographs of Microbial Keratitis. Investigative Ophthalmology & Visual Science, 66(8), 4436-4436.

Z Yang, MA Woodward, LM Niziol, M Pawar, NV Prajna, A Krishnamoorthy, Y Wang, M Lu, S Selvaraj, & S Farsiu. Self-knowledge distillation-empowered directional connectivity transformer for microbial keratitis biomarkers segmentation on slit-lamp photography. Medical Image Analysis, 102, 103533.

2023

M Chen¹, Y Wang¹, Y Shi¹, J Feng, R Feng, X Guan, … & H Wei. Brain Age Prediction Based on Quantitative Susceptibility Mapping Using the Segmentation Transformer. IEEE Journal of Biomedical and Health Informatics.

Y Wang¹, Z Li¹, J Mei¹, Z Wei¹, L Liu, C Wang, … & Y Zhou. SwinMM: Masked Multi-view with Swin Transformers for 3d Medical Image Segmentation. 2023 International Conference on Medical Image Computing and Computer-Assisted Intervention. (pp. 486-496). Cham: Springer Nature Switzerland.

Y Wang, Y Shi, H Wei. A Brain Age Estimation Network based on QSM using the Segment Transformer. 2023 International Society for Magnetic Resonance in Medicine (ISMRM).

Awards

• 2026 Oral Presentation at ARVO 2026
• 2026 Imaging Informatics Best Paper Award Runner Up at SPIE Medical Imaging 2026
• 2023 Oral Presentation at MICCAI 2023
• 2023 Outstanding Graduate of Shanghai Jiao Tong University
• 2022 Scholarship of School of Biomedical Engineering Alumni Association
• 2022 Merit Student of Shanghai Jiao Tong University
• 2021 Shanghai Municipal Government Scholarship
• 2020 Class A Scholarship of Shanghai Jiao Tong University

Service

• Conference Reviewer MICCAI 2025; MICCAI 2024;
• Journal Reviewer Image and Vision Computing; IEEE Journal of Biomedical & Health Informatics (JBHI);