TIJERE: A Novel Threat Intelligence Joint Extraction Model based on Analyst Expert Knowledge
Bibliographic record
Abstract
Extracting entities and relationships from threat intelligence reports into structured formats, such as cybersecurity knowledge graphs (CKGs), is essential for automated threat analysis, detection, and mitigation. However, existing joint extraction approaches struggle with feature confusion, language ambiguity, noise propagation, and overlapping relations, leading to low accuracy and poor model performance. This paper presents TIJERE, a novel joint entity and relation extraction framework that formulates joint extraction as a multi-sequence labeling representation (MSLR) problem, where separate sequences are generated for each entity pair. TIJERE incorporates expert domain features (EDF) to enrich positional, contextual, and semantic entity representations, improving feature distinction and classification accuracy. Additionally, SecureBERT + contextual embeddings, fine-tuned for cybersecurity text, enhance named entity recognition (NER) and relation extraction (RE) by reducing language ambiguity and improving domain-specific generalization. Empirical evaluations on the curated DNRTI-JE dataset demonstrate that TIJERE achieves state-of-the-art performance, with F1-scores exceeding 0.93 for NER and 0.98 for RE, outperforming existing methods. Additionally, this paper introduces DNRTI-JE, the first publicly available jointly labeled dataset for cybersecurity entity and relation extraction, filling a crucial gap in cyber threat intelligence automation. The dataset enables reproducible research, standardized benchmarking, and facilitates the development of next-generation AI-driven cybersecurity systems. TIJERE, along with DNRTI-JE, provides a highperformance framework for structured cybersecurity intelligence extraction, with broader applications in domains such as healthcare, finance, and bioinformatics.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.006 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.004 | 0.003 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.006 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.004 | 0.003 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".