Biography
Mathematics, learning systems, and intelligent decision-making.
I am a B.S. student in Foundational Mathematical Sciences at Dalian University of Technology, expected to graduate in June 2027. My current GPA is 4.22/5.00, with a weighted average of 91.91/100 and a major rank of 12/102.
My research centers on efficient and decision-oriented machine learning. I am currently a research intern at Peking University and the Institute of Automation, Chinese Academy of Sciences, working on biologically inspired credit assignment, decision-making LLMs, long-context modeling, and learning-based optimization.
Education
Dalian University of Technology
B.S. in Foundational Mathematical Sciences · Sep. 2023 – Jun. 2027 (expected)
- GPA
- 4.22 / 5.00
- Average
- 91.91 / 100
- Rank
- 12 / 102
Research Focus
Methods that learn efficiently and act under real constraints.
Decision-Making LLMs
Language models as policy and reasoning systems for wireless resource allocation and cooperative edge caching.
Long-Context Modeling
Linear-memory distillation and restoration methods for extending the usable context of large language models.
Publications
CAS quartiles follow the 2025 CAS upgraded list (major category); JCR quartiles follow the 2025 JCR labels. Selected conference venues also show CCF ranks, and impact factors follow official venue metrics available as of March 2026.
Global Credit Assignment via Dynamical Criticality
This ICML 2026 paper studies temporal credit assignment in recurrent, convolutional recurrent, and spiking systems through criticality-driven online local learning with constant activation memory.
Decision-Making Large Language Model for Wireless Communication: A Comprehensive Survey on Key Techniques
This survey organizes decision-making LLMs for wireless communication across data construction, architecture adaptation, reasoning control, multi-agent coordination, and open challenges.
Cooperative Edge Caching with Large Language Model in Wireless Networks
LinearARD: Linear-Memory Attention Distillation for RoPE Restoration
LinearARD aligns row-wise dense self-relations between native-RoPE teachers and RoPE-scaled students, restoring long-context performance with an exact linear-memory distillation objective.
Cluster-Aware Attention-Based Deep Reinforcement Learning for Pickup and Delivery Problems
Research Experience
Current and recent research appointments.
Research Intern
Peking University
Study temporal credit assignment beyond standard BPTT through online, biologically inspired, and memory-efficient local learning rules; analyze approximation error, stability, and scalability in recurrent and spiking systems.
Research Intern
Institute of Automation, Chinese Academy of Sciences
Work on decision-making LLMs for wireless communication, cooperative edge caching, and long-context restoration; contribute to training pipelines, experiments, visualization, technical writing, and patent materials.
Undergraduate Researcher
Dalian University of Technology
Led a neural combinatorial optimization project for pickup-and-delivery problems, including model design, reinforcement learning training, benchmark construction, comparative experiments, and manuscript writing.
Research Intern
Eastern Institute of Technology, Ningbo
Reviewed representative literature on world models and vision-language models and summarized major technical routes.
Open Source
COLA
Criticality-driven local learning for global temporal credit assignment.
OpenReviewCoopLLM-Cache
SFT and GRPO training for LLM-based cooperative edge caching.
arXivCluPDTSP
Cluster-aware deep reinforcement learning for pickup-and-delivery problems.
PaperSelected Honors
- Kaggle BirdCLEF+ 2026 Bronze Medal
- Kaggle March Machine Learning Mania 2026 Bronze Medal
- Second Prize, Climbing Cup, Dalian University of Technology
- Meritorious Winner, Mathematical Contest in Modeling
- National Second Prize, National English Translation Competition
- Academic Excellence Scholarship, Dalian University of Technology
- National First Prize, Asia and Pacific Mathematical Contest in Modeling
- Provincial First Prize, Contemporary Undergraduate Mathematical Contest in Modeling