Skip to content
View Irish-kw's full-sized avatar
💭
I may be slow to respond.
💭
I may be slow to respond.

Highlights

  • Pro

Block or report Irish-kw

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Irish-kw/README.md

About

I work on AI systems for security, graph reasoning, and observable agentic software.
My recent research focuses on recovering structure from noisy or incomplete evidence, understanding long-running APT campaigns, and building systems that make AI-agent execution behavior inspectable.


Focus On

  • CURVE — graph-theoretic reasoning under incomplete observations, with a focus on support, exhaustiveness, and occurrence relations.
  • ExecWeave — cross-layer execution provenance for observable agentic software, connecting agent semantics with processes, files, and network evidence.
  • SLM Streaming Security Memory — compact persistent memory for long-running security reasoning without replaying arbitrary discarded history.
  • GraphGeneration — conditional attack-campaign graph generation from security evidence and structured technique/campaign representations.

Featured Project — ExecWeave

ExecWeave — Cross-Layer Execution Provenance for Observable Agentic Software

A local-first observability system that combines provider-level agent semantics with operating-system runtime evidence and renders the result as an interactive execution graph.

Repository · PyPI


Publications

First-author

  • Learnable Staircase Activations for Ordinal Outputs under Mutual Information Regularization, ACM TIST (Just Accepted), 2026.
  • TGCM: Topic-Guided Generative Disentanglement of Interleaved APT Technique Sequences. arXiv preprint arXiv:2606.18651, 2026. arXiv
  • Poster: When Logs Misbehave: Retrieving Known APTs from Noisy Graphs. ACM Conference on Computer and Communications Security (CCS Poster), 2025.
  • Attention-based API locating for malware techniques. IEEE Transactions on Information Forensics and Security, 19:1199–1212, 2023.

Selected co-authored works

  • A cascade approach for APT campaign attribution in system event logs: Technique hunting and subgraph matching. ICC 2025-IEEE International Conference on Communications, 1073–1078, 2025.
  • Poster: LogCraft: Crafting CVE-Aware Synthetic Worlds (Logs). ACM Conference on Computer and Communications Security (CCS Poster), 2025.
  • Saga: Synthetic audit log generation for APT campaigns. IEEE Transactions on Dependable and Secure Computing, 2025.
  • Sparse grid imputation using unpaired imprecise auxiliary data: Theory and application to PM2.5 estimation. ACM Transactions on Knowledge Discovery from Data, 18(3):1–26, 2024.
  • Extreme event discovery with self-attention for PM2.5 anomaly prediction. IEEE Intelligent Systems, 38(2):36–45, 2023.
  • Building cybersecurity ontology for understanding and reasoning adversary tactics and techniques. 2022 IEEE International Conference on Big Data (Big Data), 4266–4274, 2022.
  • Numerical simulation for flow of rolling piston type of rotary compressor. Energies, 13(10):2526, 2020.
  • The influence factors on heat transfer performance of loop thermosyphon system. Transactions of the Canadian Society for Mechanical Engineering, 40(5):947–958, 2016.

Research Landscape

ExecWeave
Cross-layer execution provenance for observable agentic software
SLM Streaming Security Memory
Compact persistent state for long-running security reasoning
CURVE
Graph reasoning and formal guarantees under incomplete observations
GraphGeneration
Conditional generation of structured attack-campaign graphs

Observe agent and system behavior · preserve long-running evidence · reason over incomplete structure · generate campaign graphs


Research & Engineering Stack

Research and engineering stack



CUDA MITRE ATT&CK Graph ML LLM Agents

Current Focus

CURVE · ExecWeave · SLM Streaming Security Memory · GraphGeneration


Explore ExecWeave footer

Pinned Loading

  1. Attention-Based-API-Locating-for-Malware-Techniques Attention-Based-API-Locating-for-Malware-Techniques Public

    Jupyter Notebook 5 2