Incoming PhD · Computer Science & Technology

Guanchen Du 杜冠辰

Building agents that perceive, reason, and act.

See publications Learn more
Scroll
About

At the intersection of world models, embodied AI, and representation learning.

I build agents that perceive, reason, and act in complex environments.

Now
Latent world modeling and self-supervised representations for efficient planning in embodied agents.
Background
QoS prediction, graph neural networks, and recommendation — with peer-reviewed work in international journals and conferences.
Open question
How can systems learn compact, transferable representations of the world that support decision making and planning?
Long-term
General-purpose intelligent systems that learn from large-scale experience and operate across diverse real-world tasks.
By the numbers

A short record. Already shipping.

04 Selected works
Q2 Best SCI tier
4.3 Top impact factor
2026 PhD begins
News

Latest updates.

  • Starting my PhD at NJUST under Prof. Zechao Li. Soon
  • Graduating from Shantou University with a B.Eng. in Computer Science.
  • Successfully defended my undergraduate thesis on QoSDiff.
  • Co-authored paper Sparse QoS Prediction… published in Computer Communications (Elsevier).
  • GASN published in Journal of Cloud Computing (Springer).
  • BlockSys 2025 paper officially published in Springer LNCS.
  • QoSDiff preprint released on arXiv:2512.04596.
  • Launched demo.guanchendu.com.
  • Enrolled at Shantou University, College of Mathematics & Computer Science.
Education

Academic journey.

  • Ph.D. in Computer Science & Technology Nanjing University of Science & Technology Advisor: Prof. Zechao Li
  • B.Eng. in Computer Science & Technology Shantou University Advisor: Prof. Jianlong Xu
Publications

Selected work.

2026

GASN: A Graph Attention and Self-Attention Network for Blockchain Service Reliability Prediction

Guanchen Du, Jianlong Xu, Hongyu Lin, Yueming Xie.

Journal of Cloud Computing (Springer), Vol. 15, Article 50.  (CAS Q3 · SCI Q2 · IF 4.3 · 5-yr IF 4.1 · JCI 0.96 · Scopus · EI · ESI: CS)

Journal · Q2

Journal

2026

Sparse QoS Prediction for Cloud Services via Inductive Subgraph Pattern Aware Graph Neural Network

Jianlong Xu, Caiyi Chen, Qingcao Dai, Guanchen Du, Ruiqi Wang, Mingtong Li, Quanqing Guo, Yuxiang Zeng.

Computer Communications (Elsevier), Vol. 248, Article 108415.  (CAS Q3 · SCI Q2 · IF 4.3 · 5-yr IF 4.6 · JCI 0.99 · CCF-C · EI · ESI: CS)

Journal · Q2

DOI Journal

2026

An Effective Reliability Prediction Model for Blockchain Services via Hybrid Multi-layer Graph Attention and Self-attention

Guanchen Du, Jianlong Xu, Hongyu Lin.

BlockSys 2025 (Blockchain and Trustworthy Systems, Springer LNCS), pp. 410–424.  (EI · CCF-C from 2026 · Scopus)

Conference · CCF-C

Conference

2025

QoSDiff: An Implicit Topological Embedding Learning Framework Leveraging Denoising Diffusion and Adversarial Attention for Robust QoS Prediction

Guanchen Du, Jianlong Xu, Wei Wei.

Preprint, arXiv:2512.04596.

Preprint

arXiv Demo Poster Code

A full publication list is on Google Scholar.

Research

Interests.

  • World Models
  • Embodied AI
  • Foundation Models
  • Representation Learning
  • Latent Planning
  • Goal-Conditioned Control
  • AI for Services
Contact

Let's talk.

The best way to reach me is by email at 22gcdu@stu.edu.cn. Always glad to discuss research, ideas, or just say hi.