About Me
I am Zhengxu Jiang, currently an M.S. student in graph theory in the Department of Applied Mathematics, School of Mathematical Sciences, Nankai University, advised by Prof. Jiaao Li.
My current research focuses on the mechanisms of self-evolution and reinforcement learning in agents, and on how they can continually acquire new capabilities from feedback and memory, including questions of credit assignment. Alongside this work, my mathematical research lies in graph theory and combinatorial optimization. I study graph-coloring problems with special constraints, such as edge colorings involving rainbow cycles or forbidden subgraphs, as well as the use of combinatorial optimization tools in structural graph theory.
I graduated from Ocean University of China with an honors bachelor’s degree in mathematics, where I worked on dynamic modeling over complex networks, including information flow on graphs and the evolution of entity states, to address problems with practical significance. If you are interested in these areas, please feel free to contact me!
Education
Research Interests
Reinforcement Learning and Self-Evolution in Agents
Studying how agents continually acquire new capabilities from interaction feedback, memory, and experience, with a focus on credit assignment in reinforcement learning, skill learning and reuse, and joint optimization of policies and memory. Exploring structured representations for organizing experience, reasoning, and continual adaptation.
Structural Graph Theory and Combinatorial Optimization
Coloring problems with additional conditions involving cycle structures, signed graphs, and the structural properties of related graph classes.
Research Experience
Tsinghua University Natural Language Processing and Social Computing Lab (THUNLP)
Research InternResearch Focus: Knowledge memory and enhancement for large language models.
Publications and Research
B-coloring of K₂,ₜ-free planar graphs. arXiv:2609.12519 [math.CO] (2026).[paper]
(Authors are listed in strict alphabetical order, as customary in mathematics.)
Projects
Cache-aware DAG Scheduling for NPUs[code]
Models SIMD/NPU operator dependencies as a directed acyclic graph and jointly studies topological scheduling, data residency, and pipeline optimization under limited on-chip cache to reduce avoidable execution slack and data movement.
Mechanisms of the Information Cocoon Effect in Social Networks[code]
Uses a weighted directed network to model information reception, processing, diffusion, and opinion feedback, combining diffusion dynamics with structural analysis to study anomalous amplification, key-node identification, and the formation and intervention of information cocoons.
Other Competition Experience
Breaking Information Cocoons and Information-balancing Strategies[paper]
A large-scale weighted-network model using naive Bayes and ALS to study recommendation feedback and information concentration.
Graph-based Control and Optimization of Complex Water Networks[paper]
A weighted directed-graph and multi-objective model of the Great Lakes system under water-balance constraints.
Geometric Localization and Position Adjustment of UAV Swarms[paper]
Bearing-only localization and formation correction based on geometric constraints and triangulation.
Academic Activities
Summer Course on Combinatorics and Graph Theory
BeijingBUPT2022 Shenzhen Cup Award Ceremony and Symposium
ShenzhenShenzhen MSU-BIT UniversityAwards and Honors
- Shiing-Shen Chern Freshman Scholarship, Nankai University (Sep. 2025)
- Gongneng Scholarship, Nankai University (Sep. 2025)
- Graduate Recommendation Scholarship, Nankai University (Sep. 2025)
- Honorary Bachelor of Science, Ocean University of China (May 2025)
- Undergraduate thesis recognized as an Outstanding Undergraduate Thesis by Ocean University of China (May 2025)
- Bank of China Self-Improvement Scholarship (Nov. 2024)
- First-Class Comprehensive Scholarship, Ocean University of China (Nov. 2024)
- Outstanding Student Leader, Ocean University of China (Nov. 2024)
- Outstanding Communist Youth League Cadre, Ocean University of China (Apr. 2024)
- Second-Class Comprehensive Scholarship, Ocean University of China (Nov. 2023)
- National Scholarship for Undergraduate Students (Nov. 2022)
- First-Class Comprehensive Scholarship, Ocean University of China (Nov. 2022)
- Outstanding Student, Ocean University of China (Nov. 2022)
Skills
- Algorithms, optimization, and machine learning
- Graph algorithms, SAT/SMT, IP/MILP, NP-hard heuristics, and computational-graph scheduling; neural networks, CNN/RNN/LSTM, GNNs, Transformers, and HMMs; PyTorch and TensorFlow
- English
- CET-6: 517; strong academic English for reading, writing, and presentation, including English academic posters
- Tools
- SSH, Git/GitHub, Linux, LaTeX, and effective use of Codex and other AI tools
News
- Posted “B-coloring of K₂,ₜ-free planar graphs” on arXiv.
- Posted “B-coloring of grid graphs” on arXiv.
- Won a National Third Prize in the China Postgraduate Mathematical Contest in Modeling.
- Began M.S. studies in Applied Mathematics at Nankai University.
- Graduated from Ocean University of China with an Honorary B.S.; the undergraduate thesis was recognized as an Outstanding Undergraduate Thesis by the university.
- The university Undergraduate Research Development Program project “Mechanisms of the Information Cocoon Effect in Social Networks” concluded with an Excellent rating.