R²GCurL: Reinforced Robust Knowledge Tracing via Dynamic Graph Curriculum Learning
2026 · Intelligent Education
ACM TOIS 2026
I am a Research Fellow in the College of Computing and Data Science (CCDS) at Nanyang Technological University, Singapore. My postdoctoral supervisor is Professor Dacheng Tao. I received my bachelor’s degree in Computer Science and Technology in 2020 and my PhD in January 2026 from Beihang University, where Professor Wenjun Wu was my doctoral advisor.
My current research focuses on reasoning dynamics in foundation models and embodied multi-agent systems. My earlier work includes graph data mining and intelligent education.
I study how reasoning changes as information becomes available, from intermediate states in foundation models to joint plans in robot teams.
Theoretical foundation
Denoising in a diffusion language model gives me a sequence of intermediate states to examine. I want to understand when the model has enough information to answer, where errors arise, and when another step can help.
I study collective reasoning in teams of heterogeneous robots, such as a wheeled unmanned ground vehicle, a quadruped, a drone, and a robotic arm. I ask how they should share observations, divide the work, and revise their joint plan.
Joint plan

2026 · Intelligent Education
ACM TOIS 2026

2025 · Reasoning Dynamics
NeurIPS 2025 · Datasets & BenchmarksarXivCode & data

2025 · Intelligent Education
IEEE TKDE 2025

2025 · Intelligent Education
AAAI 2025arXiv

2024 · Intelligent Education
IEEE ICDE 2024arXiv

2023 · Intelligent Education
AAAI 2023arXiv

ICLR 2026arXiv

WWW 2026

ICML 2026 · ADMIRE group publication

AAAI 2026 · ADMIRE group publicationarXiv

ACL 2024

IEEE TKDE 2026

ACM KDD 2025

ACM TOIS 2024

IEEE TKDE 2022 · ADMIRE group publication

IJCAI-ECAI 2026 · ADMIRE group publication

2025 · Reasoning Dynamics

Neural Networks · January 2027 issuearXiv

2025 · Reasoning Dynamics
arXiv preprintReading list

Findings of ACL 2026arXiv

Findings of ACL 2026 · ADMIRE group publicationarXiv

arXiv 2026 · ADMIRE group publication

arXiv 2026

arXiv 2026

arXiv 2026

arXiv 2025

arXiv 2025

arXiv 2025

ICLR 2026 · Embodied · ADMIRE group publicationarXiv

IEEE TMM 2026 · Embodied · ADMIRE group publication

IEEE TMC 2025 · Embodied · ADMIRE group publication

AAAI 2024 · Embodied · ADMIRE group publicationarXiv

IEEE ICRA 2026 · Embodied · ADMIRE group publication

IEEE T-ASE 2026 · Embodied · ADMIRE group publication

IEEE/RSJ IROS 2025 · Embodied · ADMIRE group publication

IEEE ICRA 2024 · Embodied · ADMIRE group publication

IEEE/RSJ IROS 2024 · Embodied · ADMIRE group publicationarXiv

IEEE IJCNN 2025 · Embodied · ADMIRE group publicationarXiv

arXiv 2026 · Embodied · ADMIRE group publication

arXiv 2026 · Embodied · ADMIRE group publication

arXiv 2026 · Embodied · ADMIRE group publicationProject page

arXiv 2026 · Embodied · ADMIRE group publication

arXiv 2026 · Embodied · ADMIRE group publication

arXiv 2026 · Embodied

IEEE TSC 2023 · Digital · ADMIRE group publication

EMNLP 2026 · DigitalarXiv

Findings of EMNLP 2026 · DigitalarXiv

IEEE ICWS 2026 · Digital · ADMIRE group publicationarXiv

IEEE ICWS 2026 · Digital · ADMIRE group publicationarXiv

arXiv 2026 · Digital

Findings of EMNLP 2025 · DigitalarXiv

arXiv 2025 · Digital

arXiv 2025 · Digital · ADMIRE group publication

2025 · Multi-Agent Systems
arXiv preprint · DigitalCode
Best Paper Runner-Up · ICML 2025 CFAgentic Workshop · Certificate (PDF)

IEEE TNNLS 2026

IEEE JBHI 2026

IEEE TLT 2023 · ADMIRE group publication

Neurocomputing 2026arXiv

IEEE SMC 2025arXiv

Journal of Neural Engineering 2024
* Equal contribution.