About me

I am a first-year Master’s student in Computer and Information Science with a concentration in Artificial Intelligence at the University of Pennsylvania. My research focuses on advancing AI for medical applications, particularly in pathology, with the goal of assisting doctors and clinicians in improving diagnostic precision. My work spans multiple areas, including leveraging computer vision to enhance segmentation and classification in medical imaging. Beyond technical performance, I am dedicated to refining AI systems to not only function effectively but also reason critically and ethically, ensuring their impact extends meaningfully to real-world healthcare applications and beyond.

I am eager to further contribute to these areas through a Ph.D. program, where I aim to develop AI models that are not only more precise but also more interpretable, robust, and aligned with ethical considerations in healthcare.

I am fortunate to be advised by Professor Zhi Huang as a research assistant at Perelman School of Medicine at Upenn, where my work primarily focuses on innovating AI/ML models for medical image analysis. My research involves developing and optimizing algorithms and intelligent agents to improve cell nuclei-based segmentation and classification, ultimately enhancing automated pathology workflows.

📚 Education

• University of Pennsylvania, M.S. in Computer and Information Science, AI Concentration
  Aug 2024 – May 2026
  GPA: 4.0/4.0
  Relevant Coursework: Machine Learning, NLP, Deep Learning, Bayesian Optimization

• New York University, B.A. in Computer Science and Mathematics
  Feb 2021 – May 2024
  GPA: 3.88/4.0, University Honors Scholar, Dean’s List
  Relevant Coursework: Algorithms, Simulation, Data Management, Numerical/Real Analysis

🧪 Research Experience

• University of Pennsylvania — RA with Prof. Zhi Huang
  Developed NuClass, a ViT-based vision–language model for zero-shot nuclei classification using ontology-guided prompts.
  Trained on 8M+ patches across 11 datasets, achieving state-of-the-art results under low-label regimes.

• University of Pennsylvania — RA with Prof. Dan Roth
  Co-developed DeepTraceReward, a reward model benchmark for video fakeness based on fine-grained multimodal clues.
  Surpassed GPT-4.1 by 20.5% on fake clue detection.

• Capital Normal University — RA with Prof. Hongxiao Wang
  Enhanced segmentation accuracy of small abdominal organs by fine-tuning MedSAM on 8k radiology images.

• Microsoft Research NYC — RA with Dr. Jake Hofman
  Analyzed CitiBike impact using ML techniques in a real-world data science pipeline.

📄 Selected Publications

• NuClass: An ontology-driven vision–language foundation model for zero-shot nuclei classification
  CVPR 2025, MMFM-BIOMED Workshop
  → Aligns ViT features with hierarchical prompts for pathology classification.

• Learning Fine-grained Rewards on Video Generation Fakeness via Multimodal Language Models
  Under Review at NeurIPS 2025
  → Introduces DeepTraceReward benchmark; outperforms GPT-4.1 in fake clue detection.

I was a member of NYU Women in Science (Wins), a program recognizing undergraduate scholars for academic achievement and commitment to STEM inclusivity.

I was very fortunate to be advised by Prof. Nan Xu from NYU Shanghai as a Research Assistant, focusing on the dynamics of CUSMA trading. Additionally, I was advised by Dr. Terrance Chen from the NYU Department of Sociology, where I investigated the evolution and impact of ideological terms within the media. Furthermore, under the guidance of Dr. Jake Hofman, I collaborated with scientists at the Microsoft Research Lab in New York to apply machine learning techniques to assess CitiBike’s impact on NYC. Most recently, at Capital Normal University, I worked under the supervision of Prof. Hongxiao Wang, where my research focused on enhancing segmentation accuracy for small abdominal organs using machine learning models.

Eager to connect with professionals and peers in technology, research, and innovation, I welcome opportunities for collaboration and knowledge-sharing. Feel free to connect!

I am eager to find a Research Assistant position in the field of NLP, ML, or other related areas. Please feel free to contact me: xuyinuo@cis.upenn.edu.

You can download my CV here: Yinuo Xu’s Resume