Diksha Goel

Diksha Goel

Research Scientist | CSIRO's Data61

About Me

I am a Research Scientist at CSIRO’s Data61 in Melbourne, Australia, where I develop autonomous, adaptive cybersecurity systems to defend against evolving threats and protect critical infrastructure. My research combines Artificial Intelligence, Game Theory, and Graph Data Mining to build security solutions that can anticipate, detect, and respond to attacks in real time. By integrating machine learning, multi-agent systems, and graph analytics, I design defense architectures that are not only technically robust but also scalable, resilient, and aligned with real-world operational demands. My expertise spans graph-based anomaly detection, reinforcement learning for cyber threat mitigation, and strategic reasoning under uncertainty. I’m particularly focused on advancing security systems that move beyond reactive response, towards self-adaptive, intelligent defense mechanisms that can evolve with the threat landscape.

Prior to my current role, I was a Postdoctoral Fellow at CSIRO’s Data61 in collaboration with the Cyber Security Cooperative Research Centre (CSCRC), where I contributed to national cyber defense initiatives, delivering research-driven solutions for government, infrastructure operators, and industry partners.

Research

My research focuses on the design of AI-driven cybersecurity frameworks that are autonomous, scalable, and resilient, addressing the defense requirements of complex, adversarial digital systems.

For a comprehensive list of my publications, please click here.

Autonomous Cyber Defense

Development of real-time, intelligent cyber defense systems that reduce human dependence in high-stakes security operations.

  • LLM-driven threat detection and automated response generation
  • Reinforcement learning for adaptive policy learning and attack mitigation
  • Continual learning models to evolve with dynamic threat landscapes

Graph-Based Network Security

Application of graph-based learning and optimization techniques to detect anomalies and secure complex system environments.

  • Graph neural networks (GNNs) for large-scale anomaly and intrusion detection
  • Graph combinatorial optimization for vulnerability analysis and hardening
  • Modeling of system-level interactions and structural dependencies

Game-Theoretic Cybersecurity

Use of game-theoretic models to formalize attacker-defender dynamics and optimize defensive strategies under uncertainty.

  • Stackelberg and evolutionary games for adversarial modeling
  • Defense strategy optimization under uncertainty and incomplete information
  • Reinforcement learning for adaptive strategy formulation in adversarial settings
  • Application of deception and deterrence to disrupt adversarial intent

Education

Doctor of Philosophy

2019 - 2023

Master of Technology

2016 - 2018

Bachelor of Technology

2011 - 2015