Yuanyi Wang

Yuanyi Wang

The Hong Kong Polytechnic University (PolyU)

👋 About Me

I am a Ph.D. student at The Hong Kong Polytechnic University (PolyU), advised by Prof. Hongxia Yang.

My research centers on model fusion: combining trained models from different domains, architectures, or datasets into a single stronger model that aggregates complementary knowledge and improves performance, generalization, and efficiency. More broadly, I am interested in large language models, multimodal learning, and knowledge graphs.

Before joining PolyU, I received my B.S. from the College of Science at Beijing University of Posts and Telecommunications (BUPT), and was recommended for postgraduate study at the State Key Laboratory of Networking and Switching Technology, BUPT.

I am also a research intern at InfiX-AI. Our models and resources are available on Hugging Face.

🔥 News

  • May 2026 — Our paper Model Merging Scaling Laws in Large Language Models was accepted to ICML 2026.
  • 2025 — Two papers on LLM fusion were accepted to NeurIPS 2025.
  • 2025 — Our model InfiGFusion-14B was released on Hugging Face and received 4K+ stars.
  • 2024 — Our paper on multi-modal entity alignment was selected as a Best Paper Candidate at ICMR 2024.

📝 Publications

Years 2026 · 2025 · 2024

2026

ICML 2026 Representative
Model Merging Scaling Laws in Large Language Models
Yuanyi Wang, Yanggan Gu, Yiming Zhang, Qi Zhou, Zhaoyi Yan, Congkai Xie, Xinyao Wang, Jianbo Yuan, Hongxia Yang
Paper Code Repo
ICML 2026 Workshop
Access Sets Matter: Budgeting Expert Reads for Scalable Weight-Space Model Merging
Yuanyi Wang, Yanggan Gu, Su Lu, Yifan Yang, Zhaoyi Yan, Congkai Xie, Jianmin Wu, Hongxia Yang
Paper Code
arXiv Representative
Not All Disagreement Is Learnable: Token Teachability in On-Policy Distillation
Yuanyi Wang, Su Lu, Yanggan Gu, Pengkai Wang, Yifan Yang, Zhaoyi Yan, Congkai Xie, Jianmin Wu, Hongxia Yang
Paper Code
arXiv Representative
Geometry Conflict: Explaining and Controlling Forgetting in LLM Continual Post-Training
Yuanyi Wang, Yifan Yang, Su Lu, Yanggan Gu, Pengkai Wang, Wenjun Wang, Zhaoyi Yan, Congkai Xie, Jianmin Wu, Jialun Cao, Shing-Chi Cheung, Hongxia Yang
Paper Code
arXiv
MergePipe: A Budget-Aware Parameter Management System for Scalable LLM Merging
Yuanyi Wang, Yanggan Gu, Zihao Wang, Kunxi Li, Yifan Yang, Zhaoyi Yan, Congkai Xie, Jianmin Wu, Hongxia Yang
Paper Code
arXiv
FeatCal: Feature Calibration for Post-Merging Models
Yanggan Gu, Shuo Cai, Zihao Wang, Wenjun Wang, Yuanyi Wang, Pengkai Wang, Sirui Huang, Su Lu, Jianmin Wu, Hongxia Yang
Paper Code
arXiv
E-PMQ: Expert-Guided Post-Merge Quantization with Merged-Weight Anchoring
Wenjun Wang, Yanggan Gu, Shuo Cai, Yuanyi Wang, Pengkai Wang, Jianmin Wu, Hongxia Yang
Paper Code
arXiv Co-first author
Discovering Physical Directions in Weight Space: Composing Neural PDE Experts
Pengkai Wang, Pengwei Liu, Yuanyi Wang, Guanyu Chen, Xingyu Ren, Xiaolong Li, Zhongkai Hao, Yuting Kong, Qixin Zhang, Dong Ni
Paper

2025

NeurIPS 2025 Representative
InfiGFusion: Graph-on-Logits Distillation via Efficient Gromov-Wasserstein for Model Fusion
Yuanyi Wang, Zhaoyi Yan, Yiming Zhang, Qi Zhou, Yanggan Gu, Fei Wu, Hongxia Yang
Paper Model
NeurIPS 2025
InfiFPO: Implicit Model Fusion via Preference Optimization in Large Language Models
Yanggan Gu, Yuanyi Wang, Zhaoyi Yan, Yiming Zhang, Qi Zhou, Fei Wu, Hongxia Yang
Paper Model
NAACL 2025 Representative
Rethinking Smoothness for Fast and Adaptable Entity Alignment Decoding
Yuanyi Wang, Han Li, Haifeng Sun, Lei Zhang, Bo He, Wei Tang, Tianhao Yan, Qi Qi, Jingyu Wang
Paper Code
Nexus 2025
Democratizing AI Through Model Fusion: A Comprehensive Review and Future Directions
Qi Zhou, Yiming Zhang, Yanggan Gu, Yuanyi Wang, Zhijie Sang, Zhaoyi Yan, Zhen Li, Shengyu Zhang, Fei Wu, Hongxia Yang
Paper
TKDD
Understanding and Guiding Weakly Supervised Entity Alignment with Potential Isomorphism Propagation
Haifeng Sun, Yuanyi Wang, Wei Tang, et al.
Paper Code

2024

ICDE 2024 Representative
Towards Semantic Consistency: Dirichlet Energy Driven Robust Multi-Modal Entity Alignment
Yuanyi Wang, Haifeng Sun, Jiabo Wang, Jingyu Wang, Wei Tang, Qi Qi, Shaoling Sun, Jianxin Liao
Paper Code
ICDM 2024
Interdependency Matters: Graph Alignment for Multivariate Time Series Anomaly Detection
Yuanyi Wang, Haifeng Sun, Chengsen Wang, Mengde Zhu, Jingyu Wang, Wei Tang, Qi Qi, Zirui Zhuang, Jianxin Liao
Paper Code
ICMR 2024 Best Paper Candidate
Multi-modal Entity Alignment via Position-enhanced Multi-label Propagation
Wei Tang, Yuanyi Wang
Paper Code
ICSE 2024
LogExpert: Log-based Recommended Resolutions Generation using Large Language Model
Jiabo Wang, Guojun Chu, Jingyu Wang, Haifeng Sun, Qi Qi, Yuanyi Wang, Ji Qi, Jianxin Liao
Paper

🔬 Academic Service

  • Conference reviewer: ICLR, NeurIPS, ICML, WWW, ACL
  • Journal reviewer: SoftwareX, IEEE TNNLS

🎖 Honors and Awards

  • Jun. 2025 — Outstanding Thesis Award, Top 1%
  • Jun. 2025 — Outstanding Postgraduate of Beijing, Top 1%
  • Nov. 2024 — National Scholarship, Top 1%
  • Oct. 2024 — Outstanding Postgraduate, Top 5%
  • Sep. 2022 — Outstanding Graduate of Beijing, Top 5%
  • Dec. 2021 — Star of Youth Studies, awarded to 10 BUPT students each year
  • Sep. 2019 – Sep. 2021 — National Encouragement Scholarship, three times, Top 3%

📖 Education

  • Sep. 2025 – Present — Ph.D., The Hong Kong Polytechnic University, Hong Kong
  • Sep. 2022 – Jun. 2025 — Postgraduate, State Key Laboratory of Networking and Switching Technology, BUPT, Beijing
  • Sep. 2018 – Jun. 2022 — B.S., Beijing University of Posts and Telecommunications, Beijing
  • Sep. 2015 – Jun. 2018 — Xi’an Tie Yi High School, Shaanxi

💻 Internships

  • Mar. 2025 – Present — InfiX-AI, Guangdong, China
  • Nov. 2023 – Jan. 2025 — Huawei 2012 Lab, Guangdong, China

📟 Media & Reports

🏃 Outside Research

I was a member of BUPT’s varsity team and ranked 5th in the Beijing amateur division for hammer throw. I also enjoy basketball, table tennis, badminton, and strength training.