AMLUCS 2025
Brought to you by:
Sponsored by
Supported by
Day 1
Adaptive by Design: Contextual Reinforcement Learning for Mission-Ready Cyber Defence
Jake Thomas, AdvaiAutonomous Cyber Resilience: An Infrastructure-as-Code Approach for Coalition Military Networks
Dr Konrad Wrona, NATO Communications and Information AgencyDeploying Autonomous Cyber Defence onto a Military Relevant Physical System
Alec Wilson, BMTPanel: Operationalising AI - Integrating Machine Learning into Defence Cybersecurity Systems
Helen, Dstl, with Ministry of Defence stakeholders.Autonomous Defensive Cyber
Wayne & Lara, DstlSimulation-to-Reality Gap via RL-trained ROSbots
Dr Jack Smith, AwerianTopological Extensions for Reinforcement Learning Agents (TERLA).
Tim Dudman, Riskaware
AMLUCS 2025
Sponsored by: Awerian
Day 2
RAGING MINOTAUR: Improving Defence against AI-driven Cyber-Attacks by Designing More Capable Autonomous Cyber Training Adversaries
Althea & Sharaz, DstlLessons Learned in the Application of Reinforcement Learning Agents for APT Attack Path Generation
Chad Caison, Six24 Cyber Labs, SIEGE team, DARPA CASTLEMitigating data poisoning attacks against RL from Human Feedback (RLHF) finetuning
Dr James Titchener, RaytheonText2VLM: Adapting Text-Only Datasets to Evaluate Alignment Training in Visual Language Models
Gabriel Downer, AdvaiA Statistical Pipeline for Uncertainty Quantification in ACD Test and Evaluation
Dr Miriam Apsley & Dr Alessio Zakaria, Smith InstituteModelling adversarial behaviour to enable AI predictions analogous to counterfactual reasoning
Dr Louis Gauntlett, Frazer-Nash Consultancy