File size: 5,481 Bytes
a143872
 
 
 
 
e233ac2
 
e70f16f
a143872
 
e70f16f
a143872
 
 
e233ac2
 
 
a143872
e233ac2
 
13548c0
 
 
 
e233ac2
a143872
 
 
13548c0
 
 
a143872
e233ac2
 
 
a143872
e233ac2
 
 
 
 
 
 
 
0c981af
e233ac2
a143872
e233ac2
a143872
e233ac2
a143872
e233ac2
 
 
 
a143872
e233ac2
 
 
 
a143872
 
 
 
 
 
 
d59be39
e70f16f
a143872
 
 
 
e233ac2
 
 
 
 
 
 
 
 
 
 
a143872
 
 
 
 
e233ac2
 
a143872
e233ac2
 
 
 
 
a143872
 
 
 
 
13548c0
 
 
62b86ec
13548c0
 
e233ac2
13548c0
e233ac2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
---
license: mit
language:
- en
tags:
- cybersecurity
- vulnerability
- mitre-attack
- text-classification
- fine-tuned
- securebert
base_model: ehsanaghaei/SecureBERT
---

# SecureBERT — CVE-LMTune ATT&CK Classifier (Flat)

<div align="center" style="display:inline-flex; gap:18px; align-items:center; flex-wrap:nowrap;"> <img src="https://upload.wikimedia.org/wikipedia/commons/thumb/5/5b/Logo_Universit%C3%A9_de_Lorraine.svg/1280px-Logo_Universit%C3%A9_de_Lorraine.svg.png" alt="Universite de Lorraine" style="height:50px; width:auto;" /> <img src="https://upload.wikimedia.org/wikipedia/commons/thumb/9/95/Inr_logo_rouge.svg/1280px-Inr_logo_rouge.svg.png" alt="INRIA" style="height:50px; width:auto;" /> <img src="https://upload.wikimedia.org/wikipedia/fr/6/6e/Logo_loria_abrege_couleur.png" alt="LORIA" style="height:70px; width:auto;" /> <img src="https://www.pepr-cybersecurite.fr/wp-content/uploads/2023/09/pep-cybersecurite-550x250-1.png" alt="SuperViZ" style="height:70px; width:auto;" /> </div>

[![GitHub](https://img.shields.io/badge/GitHub-CVE--LMTune-black?logo=github)](https://github.com/terranovafr/CVE-LMTune)
[![Paper](https://img.shields.io/badge/Paper-HAL-green?logo=information&logoColor=white)](https://hal.science/hal-05500820)
[![PhD theses.fr](https://img.shields.io/badge/Project-theses.fr-orange?logo=university&logoColor=white)](https://theses.fr/s371241)
[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
[![Zenodo Data](https://img.shields.io/badge/Zenodo-Data%20Repository-lightblue?logo=information&logoColor=white)](https://doi.org/10.5281/zenodo.16936476)

Part of the **CVE-LMTune** model suite, a collection of language models fine-tuned for multi-taxonomy vulnerability classification across widely used cybersecurity taxonomies, including CWE, CAPEC, and MITRE ATT&CK.

## Paper

> Franco Terranova, Sana Rekbi, Abdelkader Lahmadi, Isabelle Chrisment.
> *Multi-Taxonomy Vulnerability Classification with Hierarchically Finetuned Language Models.*
> The 23rd Conference on Detection of Intrusions and Malware & Vulnerability Assessment **(DIMVA '26)**.

## Overview

This model performs **multi-label ATT&CK classification** from vulnerability descriptions. Given a CVE-style description, it predicts one or more ATT&CK identifiers associated with the described vulnerability.

| Property | Value |
|----------|-------|
| Taxonomy | MITRE ATT&CK Enterprise Subtechniques |
| Task | Multi-label text classification |
| Input | Vulnerability description (e.g., CVE summary) |
| Output | One or more ATT&CK identifiers |
| Number of labels | 175 |
| Number of samples | 231,009 |
| Latest CVE update included | 17/06/2026 |
| Split | train (60%), val (20%), test (20%) |

## Evaluation Results

The model was evaluated on the held-out test set using standard multi-label classification metrics using sigmoid activation and a default threshold of 0.5.

**Ranking Metrics**
| LRAP | MRR | Coverage Error | Label Ranking Loss | P@1 | P@3 | P@5 | R@1 | R@3 | R@5 |
|------|-----|----------------|--------------------|-----|-----|-----|-----|-----|-----|
| 0.9152 | 0.9460 | 18.79 | 0.0173 | 0.9321 | 0.9084 | 0.8458 | 0.1286 | 0.3779 | 0.5554 |

**Threshold = 0.5**
| Micro P | Micro R | Micro F1 | Macro F1 | Weighted F1 | Hamming Loss | Subset Accuracy |
|--------|--------|----------|----------|------------|--------------|----------------|
| 0.8612 | 0.7767 | 0.8168 | 0.4286 | 0.8093 | 0.0264 | 0.6874 |

## Quick Start

```python
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tokenizer = AutoTokenizer.from_pretrained("Sana9/securebert-vuln2attack-flat", use_fast=False)
model = AutoModelForSequenceClassification.from_pretrained("Sana9/securebert-vuln2attack-flat")

text = "Buffer overflow vulnerability in OpenSSL allows remote attackers to execute arbitrary code."

with torch.no_grad():
    probs = torch.sigmoid(
        model(**tokenizer(text, return_tensors="pt", truncation=True)).logits
    )[0]

predictions = {
    model.config.id2label[i]: p.item()
    for i, p in enumerate(probs)
    if p > 0.5
}

print(predictions)
```

## Citation

```bibtex
@inproceedings{terranova2026multitaxonomy,
  author    = {Franco Terranova and Sana Rekbi and Abdelkader Lahmadi and Isabelle Chrisment},
  title     = {Multi-Taxonomy Vulnerability Classification with Hierarchically Finetuned Language Models},
  booktitle = {Proceedings of the International Conference on Detection of Intrusions and Malware, and Vulnerability Assessment (DIMVA)},
  year      = {2026},
  month     = jul,
  address   = {Chania, Crete, Greece},
  note      = {HAL identifier: hal-05500820v2}
}
```

## Related Resources

- 🤗 [Full model suite on Hugging Face](https://huggingface.co/Sana9)
- 💻 [CVE-LMTune — Training code (GitHub)](https://github.com/terranovafr/CVE-LMTune)
- 📦 [Zenodo — Data repository](https://doi.org/10.5281/zenodo.16936476)

## Disclaimers

- This product is a result of the use of the NVD API but is not endorsed or certified by the NVD. The same for the CVE2CAPEC project and the Hugging Face API.
- This project relies on data publicly available from the CWE, CAPEC, and MITRE ATT&CK projects.
- This work has been partially supported by the French National Research Agency under the France 2030 label (Superviz ANR-22-PECY-0008). The views reflected herein do not necessarily reflect the opinion of the French government.