Metraipolis › archive › dc_bb905415ad2f

Retrieval-Constrained Policy Optimization for Attack Technique Extraction from Cyber Threat Intelligence — arXiv:2608.06778

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.06778
License (as archived)
CC0-1.0
Archived
2026-10-19T09:47:07.304Z
SHA-256 of the archived text
ab68b14317dc7d31ace47035f17eb4b31e1240cd55c5e9e15df9dcc4689349c4
Stable URL of this page
https://metraipolis.com/documento/dc_bb905415ad2f

How to cite

Retrieval-Constrained Policy Optimization for Attack Technique Extraction from Cyber Threat Intelligence — arXiv:2608.06778 — arXiv (en), https://arxiv.org/abs/2608.06778 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bb905415ad2f (Metraipolis archive dc_bb905415ad2f, sha256 ab68b14317dc7d31ace47035f17eb4b31e1240cd55c5e9e15df9dcc4689349c4)
BibTeX
@misc{dcbb905415ad2f,
  title = {Retrieval-Constrained Policy Optimization for Attack Technique Extraction from Cyber Threat Intelligence — arXiv:2608.06778},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.06778},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bb905415ad2f; sha256 ab68b14317dc7d31ace47035f17eb4b31e1240cd55c5e9e15df9dcc4689349c4},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.06778. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Mapping cyber threat intelligence (CTI) text to MITRE ATT&CK techniques is essential for structured threat analysis, yet manual annotation is costly and does not scale. The ATT&CK taxonomy comprises several hundred attack techniques, and a single CTI passage may describe multiple techniques, making accurate and complete extraction challenging. Existing automated approaches fall short in different ways: multi-label classifiers struggle with severe class imbalance and the large label space, while LLM-based methods--retrieval pipelines and fine-tuned generators--optimize token-level objectives that treat technique annotation as sequence generation rather than set prediction, lacking direct supervision on whether the predicted technique set is correct and complete. We propose TTP-R1, a two-stage framework that combines retrieval-augmented supervised fine-tuning (SFT) with reinforcement learning using verifiable rewards (RLVR). A hybrid retriever first narrows the large label space to a candidate set, and a fine-tuned LLM learns to select the correct techniques. We then apply Group Relative Policy Optimization with a decomposed reward that directly supervises the precision, recall, and output format of the predicted technique set. Across four CTI benchmarks, TTP-R1 achieves the best average F1, improving sub-technique-level F1 by 7.4 percentage points over Claude Sonnet 4.5 with retrieval augmentation, while running 28x faster when served as an 8B-parameter model on a single GPU.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-10 in cs.CL).
Categorie arXiv: Cryptography and Security (cs.CR) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.06778 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-19T09:46:37.263Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-10?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).