Academic Collaborations¶

TI Mindmap HUB is committed to advancing research in the intersection of Generative AI and Cyber Threat Intelligence. This document details our academic collaborations, supervised research, and partnership opportunities.

Supervised Research¶

STIX 2.1 Generator — Giulio Triggiani (2024-2025)¶

Master's Thesis: Automatic Generation of STIX 2.1 Bundles using Large Language Models


Author	Giulio Triggiani
Type	Master's Thesis
University	University of Salerno (UNISA)
Supervisors	Assoc. Prof. Arcangelo Castiglione, Ing. Antonio Formato
Status	✅ Completed and Integrated
Repository	GiulioTriggiani/STIX-2.1-Generator

Abstract¶

This thesis explores the automatic generation of STIX 2.1 bundles from unstructured threat intelligence reports using Large Language Models. The research addresses the challenge of transforming narrative threat reports into structured, machine-readable formats that can be ingested by security tools.

Key Contributions¶

LLM-Based Extraction Pipeline
Extraction of STIX Domain Objects (SDOs): Threat Actors, Malware, Campaigns, Attack Patterns
Extraction of STIX Cyber Observables (SCOs): IPs, Domains, URLs, File Hashes
Automated relationship mapping using semantic analysis
Evaluation Framework
Standard evaluation: object-by-object precision/recall
Graph-based evaluation: structural similarity metrics
Comparison across multiple LLM configurations
Production Integration
Methodology integrated into TI Mindmap HUB
Processing ~40 threat reports weekly
Real-time STIX validation and export

Results¶

Metric	Result
SDO Precision	~85%
SDO Recall	~78%
Relationship Accuracy	~72%
Schema Compliance	100%

Resources¶

Citation¶

@mastersthesis{triggiani2025stix,
  title = {Automatic Generation of STIX 2.1 Bundles using Large Language Models},
  author = {Triggiani, Giulio},
  year = {2025},
  school = {University of Salerno},
  type = {Master's Thesis},
  supervisor = {Castiglione, Arcangelo and Formato, Antonio},
  url = {https://github.com/GiulioTriggiani/STIX-2.1-Generator}
}

Research Areas¶

We are actively seeking collaborations in the following areas:

1. STIX 2.1 Generation and Validation¶

Improving extraction accuracy for complex threat reports
Multi-language support for non-English reports
Automated validation and correction of generated bundles
Comparison of different LLM architectures

2. Knowledge Graph Construction¶

Building threat intelligence knowledge graphs
Entity resolution across multiple reports
Temporal analysis of threat evolution
Graph-based threat hunting queries

3. LLM Evaluation in Security Contexts¶

Benchmarking LLMs for CTI tasks
Hallucination detection in security outputs
Adversarial testing of extraction pipelines
Human-AI collaboration studies

4. MCP Protocol Applications¶

Novel MCP tool development for security
Multi-agent systems for threat analysis
Real-time intelligence sharing protocols

Collaboration Opportunities¶

Thesis Supervision¶

We offer thesis supervision for Master's and PhD students working on:

Generative AI applications in cybersecurity
Automated threat intelligence processing
STIX/TAXII implementations
Security data visualization

What we provide: - Access to TI Mindmap HUB platform and APIs - Real-world datasets (anonymized) - Technical mentorship - Potential for production integration

Joint Research Projects¶

Open to collaboration on:

Grant proposals (EU Horizon, national funding)
Joint publications
Dataset creation and sharing
Tool development

Dataset Access¶

Researchers can request access to:

Processed threat report corpus
STIX bundle samples
Evaluation datasets
API access for research purposes

How to Apply¶

For Thesis Students¶

Review our research areas
Prepare a 1-page research proposal
Contact us at info@ti-mindmap-hub.com with:
Your proposal
CV/Resume
Academic supervisor contact
Expected timeline

For Researchers¶

Contact us at info@ti-mindmap-hub.com
Describe your research interest
Propose a collaboration format

Publications¶

This section will be updated as research outputs are published.

Planned Publications¶

Evaluation of LLM accuracy in STIX generation (in preparation)
TI Mindmap HUB: Platform architecture and lessons learned (planned)

Last updated: January 2025