Yifan Feng

Yifan Feng

Postdoctoral Researcher · Postdoctoral Innovation Talent Support Program
Tsinghua University · School of Software
AI for ScienceHigh-Order Correlation ComputingHigh-Order Pattern Discovery

Making high-order patterns at every scale computable, discoverable, and designable Understand physics, design molecules, decode life

Education
Jul. 2025 —
Postdoctoral Researcher, School of Software, Tsinghua University
Supervisor: Yue Gao
Sep. 2021 — Jun. 2025
Ph.D. in Software Engineering, Tsinghua University
Supervisor: Yue Gao
Sep. 2018 — Jun. 2021
M.Eng. in Intelligence Science and Technology, Xiamen University
Supervisor: Rongrong Ji
Sep. 2014 — Jun. 2018
B.Eng. in Computer Science and Technology, Xidian University
About Me

I lead projects including an NSFC General Program, an NSFC Young Student Basic Research Program (PhD Candidate), and a subproject of a national major S&T program. I have published 40+ papers in Nature Communications, IEEE TPAMI (14 papers), ICLR, etc., with 7500+ Google Scholar citations; my most cited first-author paper has been cited 3000+ times. I led the development of DeepHypergraph, the first hypergraph computation toolbox, and Hyper-Extract, a hypergraph knowledge reasoning toolkit that ranked 2nd on GitHub Trending.

My research centers on high-order correlation computing and pattern discovery. The principles of complex systems often lie not in individual elements, but in how a group of elements co-occur and co-configure — such higher-order correlations are what I make computable. Using hypergraphs as a new tool, I build modeling theory, algorithms, and systems for high-order correlations, and characterize their capability boundaries.

Looking ahead, I am working toward autonomous scientific discovery driven by high-order correlations, from physical intelligence and protein design to tumor microenvironment analysis. Feel free to reach out if you are interested in these directions.

I am currently on the job market for faculty positions, mainly in Beijing — any information or referrals would be greatly appreciated.

News

2022 — 2026
2026
Sep. 2026 Survey "Hypergraph Neural Networks: Theoretical Foundations and Applications" accepted by IEEE TPAMI. IEEE TPAMI
Sep. 2026 Hyper-Extract surpassed 4K+ Stars with the v0.10 series release (JSON-LD & Cypher exports, Gemini, MCP parity, Scoped Chat). 4K+ Stars
Aug. 2026 Awarded an NSFC General Program grant as Principal Investigator. NSFC
Jun. 2026 Selected for the Postdoctoral Innovation Talent Support Program (BX Program). BX Program
Jun. 2026 Hyper-Extract ranked #2 on GitHub Trending worldwide. Trending #2
Jun. 2026 Paper "How Powerful are Hypergraph Neural Networks?" accepted by IEEE TPAMI. IEEE TPAMI
Jun. 2026 Hyper-KGGen received an Oral presentation at KDD 2026. Oral
Apr. 2026 Released Hyper-Extract, an open-source framework for general knowledge extraction over cross-domain, multi-type data structures. Open Source
Mar. 2026 Paper "Hyper-RAG" accepted by Nature Communications. Nature Portfolio
Mar. 2026 Paper "Knowledge-Embedded Hypergraph Neural Networks" accepted by IEEE TPAMI. IEEE TPAMI
Jan. 2026 Doctoral dissertation nominated for the CSIG Doctoral Dissertation Incentive Program. Award
2025
Dec. 2025 Three papers (Hypergraph Foundation Model, HGNN Shield, HGNNv2) accepted by IEEE TPAMI. IEEE TPAMI ×3
Nov. 2025 Paper Cog-RAG accepted by AAAI 2026. AAAI
Sep. 2025 Listed among the World's Top 2% Scientists (2024 & 2025). Award
Sep. 2025 Paper "Reinterpreting Hypergraph Kernels" accepted by IEEE TPAMI. IEEE TPAMI
Jul. 2025 Selected as a Shuimu Tsinghua Scholar. Award
Jul. 2025 Paper "Self-Supervised Hypergraph Training Framework" accepted by IEEE TPAMI. IEEE TPAMI
Jun. 2025 Named Tsinghua University Academic Rising Star (10 recipients university-wide). Award
Jun. 2025 Named Beijing Outstanding Graduate. Award
Jun. 2025 Paper "Kernelized Hypergraph Neural Networks" accepted by IEEE TPAMI. IEEE TPAMI
Jun. 2025 Paper Hyper-Depth accepted by ICCV 2025. ICCV
Mar. 2025 Paper "Hypergraph Foundation Model for Brain Disease Diagnosis" accepted by IEEE TNNLS. IEEE TNNLS
Feb. 2025 Paper "Mode Hypergraph Neural Network" accepted by IEEE TNNLS. IEEE TNNLS
Jan. 2025 Paper "Beyond Graphs" accepted by ICLR 2025. ICLR
Jan. 2025 Paper "Residual Fuzzy Alignment on Hypergraph" accepted by IEEE TMM. IEEE TMM
2024
Dec. 2024 Two papers (Hyper-YOLO, Cross-Modal 3D Shape Retrieval) accepted by IEEE TPAMI. IEEE TPAMI ×2
Nov. 2024 Paper "Hypergraph-Based Remaining Prototype Alignment" accepted by IEEE TMM. IEEE TMM
Dec. 2024 Received the First-class Comprehensive Excellence Scholarship, Tsinghua University (third consecutive year, 2022–2024). Award
Sep. 2024 Three papers accepted by NeurIPS 2024. NeurIPS ×3
Jul. 2024 Received the ICME 2024 Best Student Paper Award. Award
May. 2024 Awarded the NSFC Young Student Basic Research Program (PhD Candidate), first cohort. NSFC
Apr. 2024 Won the Silver Medal at the Geneva International Exhibition of Inventions (student first). Award
Apr. 2024 One paper accepted by IJCAI 2024. IJCAI
Apr. 2024 One paper accepted by IEEE TIP. IEEE TIP
Feb. 2024 v0.9.4 of DHG released: 6 new hypergraph datasets and bug fixes. Release
Jan. 2024 Two papers accepted by ICLR 2024. ICLR
2023
Dec. 2023 Named Person of the Year 2023 (Research Innovation), School of Software, Tsinghua University. Award
Dec. 2023 One paper accepted by IEEE TPAMI. IEEE TPAMI
Nov. 2023 One paper accepted by IEEE TPAMI. IEEE TPAMI
Oct. 2023 Invited talk at PRCV 2023: Hypergraph Computation — Methods and Applications. Talk
Oct. 2023 One paper accepted by IEEE TNNLS. IEEE TNNLS
2022
Dec. 2022 v0.9.3 of DHG released: a new hypergraph dataset, bug fixes, and new hypergraph operations. Release
Sep. 2022 v0.9.2 of DHG released: 21 datasets, 6 SOTA methods, and structure/feature visualizations. Release
Sep. 2022 One paper accepted by IEEE TPAMI. IEEE TPAMI
Sep. 2022 Talk at 3DOR 2022 on Open-Set 3D Object Retrieval. Talk
Aug. 2022 First version v0.9.1 of DeepHypergraph (DHG) released. Release
Jul. 2022 One paper accepted by Computers & Graphics. C&G
Jun. 2022 Built an online multi-modal 3D object retrieval system — welcome to play. Demo
May. 2022 One paper accepted by IEEE TPAMI. IEEE TPAMI
Feb. 2022 Organizing the track "Open-Set 3D Object Retrieval" at SHREC'22. Track

Selected Publications

Hypergraph
Hyper-Fold
Hyper-Fold: Exploring the Expressive Limit of Sequence-Geometry Learning for Proteins via Hypergraph Modeling
Yifan Feng, Guanjie Cheng, Shihui Ying, Shaoyi Du, Yue Gao
arXiv · 2026
PDFCode
Hyper-RAG
Hyper-RAG: Combating LLM Hallucinations using Hypergraph-Driven Retrieval-Augmented Generation
Yifan Feng#, Hao Hu#, Shihui Ying#, Xingliang Hou, Shiquan Liu, Mingyuan Yang, Junchang Li, Shaoyi Du, Nanning Zheng, Han Hu, Yue Gao
Nature Communications · 2026 Nature Portfolio Journal
PDFCode
Hyper-YOLO
Hyper-YOLO: When Visual Object Detection Meets Hypergraph Computation
Yifan Feng, Jiangang Huang, Shaoyi Du, Shihui Ying, Xu Zhuang, Jason Wang, Jun-Hai Yong, Guiguang Ding, Rongrong Ji, Yue Gao
IEEE TPAMI · 2025 500+ Citations
PDFCode
HGNN
Hypergraph Neural Networks
Yifan Feng, Haoxuan You, Zizhao Zhang, Rongrong Ji, Yue Gao
AAAI · 2019 Oral · 3000+ Citations
PDFCode
View All →
11,646Total visits · 53 countries/regionsUpdated every 5 minutes
China5,015United States1,911Singapore1,070Hong Kong, China1,069Japan472United Kingdom267Germany226South Korea194Australia164Canada157France145India134Netherlands106Switzerland103Italy74