Yiyao Zhang
Expert Decision Systems  ·  Multi-Agent Reinforcement Learning  ·  Cyber Defence
Adelaide, South Australia, Australia yiyao663@gmail.com Google Scholar LinkedIn
Research Statement
My research develops trustworthy autonomous decision systems that unite causal reasoning, reinforcement learning, neuro-symbolic inference, and multi-agent coordination. I focus on AI architectures that reason strategically under uncertainty, justify their decisions, and remain reliable in high-stakes environments such as cyber defence, financial systems, and scientific decision support. My long-term agenda is to build verifiable, aligned, and safety-aware intelligent agents that integrate learning, structured reasoning, and formal reliability guarantees into deployable decision-support systems.
3
Journal Articles
2
Conference Papers
6
Under Review
15.5
Top Impact Factor
A*
Top Venue Target
Education
University of Wollongong, Wollongong, Australia
2026–2030
PhD Offer in Computer Science
Advisor: Prof Jun Shen
University of South Carolina, Columbia, USA
2017–2020
PhD-level Study in Statistics
PhD Qualifying Examination Passed
Completed PhD-level coursework and qualifying examination; program interrupted during the COVID-19 period.
National University of Singapore, Singapore
—
PhD offer in Mathematics
Adelaide University, Adelaide, Australia
2024–2026
Master of Data Science
GPA: 6.4/7.0
University of South Carolina, Columbia, USA
2017–2019
Master-level Data Science and Statistics Training
GPA: 3.6/4.0
University of Melbourne; UNSW; NYU Courant; Georgia Tech
—
Master offers in data science and mathematics
Jilin University, Changchun, China
2011–2015
Bachelor of Science in Computational Mathematics
Aoqing Tang Honor Class  ·  GPA: 93/100
Jilin University is a Project 985/211 institution in China; selected into an elite mathematics honours program.
University of California, Los Angeles, USA
2014
Bachelor Summer Research Program
UCLA-CSST Program  ·  GPA: 3.8/4.0
Georgia Institute of Technology, Atlanta, USA
2014
Bachelor Exchange Study in Applied Mathematics
GPA: 3.7/4.0
Selected Publications
Full list available on Google Scholar →
Journal Articles
  1. [1]
    Explainable Autonomous Cyber Defense using Adversarial Multi-Agent Reinforcement Learning
    Y. Zhang, D. Goel, and H. Ahmad
    Expert Systems with Applications, accepted, No. ESWA-D-26-05576R2, 2026
    ESWA IF: 7.5 2026
  2. [2]
    RegimeFolio: A Regime-Aware ML System for Sectoral Portfolio Optimization in Dynamic Markets
    Y. Zhang, D. Goel, H. Ahmad, and C. Szabo
    IEEE Access, Vol. 13, pp. 184722–184744, 2025
    IEEE Access IF: 3.6 2025
  3. [3]
    AgentVM: Agentic Multi-Agent Software Vulnerability Management
    A. Arifin, H. Ahmad, Y. Zhang, and D. Goel
    IEEE Software, accepted, No. SWSI-2026-01-0012.R2, 2026
    IEEE Software IF: 3.0 2026
Conference Papers
  1. [4]
    Synergistic MARL: Unifying Physics-Informed, Meta Learning, and Hybrid Learning
    Y. Zhang, Y. Dong, K. Yu, W. Liu, J. Zhou, B. Liang, and F. Chen
    Australasian Joint Conference on Artificial Intelligence (AJCAI), 2025
    AJCAI CORE Rank B 2025
  2. [5]
    AI-Edu-App: An Integrated AI Platform for K–12 Learning
    Y. Zhang, J. Huang, et al.
    ACM International Conference on Education Technology and AI (ICETAI), 2025
    ACM ICETAI 2025
Manuscripts Under Review
  1. [6]
    VACS: Value-Aligned Compositional Shielding for Multi-Agent Reasoning
    Y. Zhang, D. Goel, H. Ahmad, H. Li, and J. Shen
    Under Review — Information Fusion
    Under Review IF: 15.5
  2. [7]
    CAUSALNAV: Neuro-Symbolic Reasoning over Learned Causal World Models
    Y. Zhang, D. Goel, H. Ahmad, H. Li, and J. Shen
    Under Review — NeurIPS 2026
    Under Review CORE A*
  3. [8]
    Debate-to-Act: Dependable Multi-Agent Cyber Defense under Partial Observability and Safety Constraints
    Y. Zhang, D. Goel, H. Ahmad, H. Li, and J. Shen
    Under Review — IEEE Transactions on Dependable and Secure Computing (TDSC)
    Under Review IF: 7.5
  4. [9]
    Beyond Correctness: Decomposing LLM Rewards into Grounded Multi-Signal Channels for Step-Level Verification
    Y. Zhang, D. Goel, H. Ahmad, H. Li, and J. Shen
    Under Review — ACM Transactions on Intelligent Systems and Technology (TIST)
    Under Review IF: 6.6
  5. [10]
    SAGE: A Predict-Then-Decide World Model Framework for Autonomous Cyber Defense
    Y. Zhang, D. Goel, H. Ahmad, and M. Guo
    Under Review — Journal of Information Security and Applications (JISA)
    Under Review IF: 3.7
  6. [11]
    AdapTutor: An Evolving Dual-Agent Framework for Synchronous Learner Support in Online Learning Environments
    Y. Zhang, X. Fan, and L. Yue
    Under Review — IEEE Transactions on Education
    Under Review IF: 2.0
Selected Projects
Explainable Autonomous Cyber Defence with Adversarial MARL
2025–2026
Collaboration with Diksha Goel and Hussain Ahmad
→ Expert Systems with Applications
  • Developed a causal multi-agent decision framework for autonomous cyber defence under adversarial uncertainty and ambiguous telemetry.
  • Integrated causal modelling, adversarial dual-policy control, and explainable response reasoning to improve robust cyber decision-making.
Agentic AI for Software Vulnerability Management
2025–2026
Collaboration with Asrul Arifin, Hussain Ahmad, and Diksha Goel
→ IEEE Software
  • Designed an agentic multi-agent vulnerability management system for aggregating, enriching, prioritising, and reporting software security findings.
  • Combined LLM-based agents with security tools to reduce raw scanner outputs into compact, actionable vulnerability queues.
Regime-Aware Machine Learning for Portfolio Optimisation
2025
Collaboration with Diksha Goel, Hussain Ahmad, and Claudia Szabo
→ IEEE Access
  • Developed RegimeFolio, a regime-aware ML framework for sector-specialised portfolio optimisation in dynamic markets.
  • Integrated volatility-regime detection, sector-specific ensemble forecasting, and adaptive mean–variance allocation.
Research & Professional Experience
Research Intern — AI for Financial Markets
2024–2025
Australian Institute for Machine Learning, Adelaide, Australia
  • Investigated deep learning and ML techniques for algorithmic trading and market decision support.
  • Designed, implemented, and backtested trading strategies using PyTorch, Scikit-learn, Pandas, and NumPy.
  • Collaborated with PhD researchers to adapt advanced AI models for market pattern identification and risk-aware decision-making.
AI and Software Developer
2024–2025
Upwork–Syncove, Adelaide, Australia
  • Designed quantitative analysis tools for financial markets using Python, Pandas, and NumPy.
  • Developed Advanced Performance Analyzer Pro, a Python desktop application for probabilistic recommendation and performance analytics.
  • Implemented risk-optimised allocation algorithms based on Modern Portfolio Theory and decision analytics.
Algorithm Engineer
2020–2023
Haier Electronics, Qingdao, China
  • Developed ML algorithms for sales-volume forecasting, improving forecasting accuracy by approximately 15%.
  • Applied NLP methods to customer feedback to extract product-improvement insights and decision signals.
  • Designed and maintained database systems for large-scale sales analytics and operational decision support.
Teaching Assistant
2017–2020
University of South Carolina, Columbia, USA
  • Taught statistical concepts, data analysis, and modelling fundamentals using R and Python.
  • Supported undergraduate tutorials in statistics, ML foundations, and applied data analysis.
Research Project Team Member
2012–2015
Jilin University, Changchun, China
  • Developed generalised linear models and predictive frameworks for large environmental datasets.
  • Conducted statistical analysis using R and SPSS to identify patterns in complex environmental and public-health data.
Referees
Prof. Ran Zhang
Vice President & Dean, School of Mathematics, Jilin University
Director, Tianyuan Mathematical Center in Northeast China
zhangran@jlu.edu.cn
+86 135 0447 5032
Prof. Jun Shen
Head, School of Computing & IT, University of Wollongong
Director, Center for Applied Computing
jun_shen@uow.edu.au
+61 2 4221 3873
Prof. Lin Liu
Director, 4LLab Data Analytics Group
Professor, STEM Unit, Adelaide University
Lin.Liu@adelaide.edu.au
Prof. Wotao Yin
Director, Decision Intelligence Lab, DAMO Academy, Alibaba
Professor, University of California, Los Angeles
wotao.yin@alibaba-inc.com