0% Complete
فارسی
Home
/
شانزدهمین کنفرانس بین المللی فناوری اطلاعات و دانش
Robustness Gap in NLP Models for Vulnerability Descriptions: Benchmarking and Data Augmentation
Authors :
AmirHossein Majd
1
Mahdi Yousefikia
2
Saghar Ghasemzadeh
3
Amirreza Asari
4
Arya Khoshnavataher
5
Seyedeh Leili Mirtaheri
6
1- University of Calabria
2- دانشگاه خوارزمی
3- دانشگاه خوارزمی
4- دانشگاه خوارزمی
5- دانشگاه خوارزمی
6- University of Calabria
Keywords :
Software Vulnerabilities،Natural Language Processing،Robustness Benchmark،Noise Injection،Exploitability Prediction،Data Augmentation،Cybersecurity
Abstract :
Software vulnerability descriptions from CVE/NVD are the primary corpus for analysis, prioritization, and risk management in cybersecurity. Yet natural noise (typos, synonym substitutions, lexical variety) and adversarial perturbations undermine the accuracy and trustworthiness of NLP models. This paper presents, to our knowledge, the first systematic benchmark of NLP robustness on vulnerability descriptions. We train nine diverse architectures—lightweight transformers (MiniLM, MPNet, SBERT), hybrid models (BERT-LSTM, TextRCNN), and classical recurrent networks (BiLSTM, LSTM)—on a balanced dataset of over 56,000 real-world records from NVD and Exploit-DB, and fine-tune them for exploitability prediction. For comprehensive evaluation, we inject three noise families into test sets at levels from 10% to 80%: character-level edits (substitutions/swaps), synonym replacements using WordNet, and composite adversarial attacks generated with TextAttack. Performance declines across all models as noise rises, but vulnerability profiles differ: MiniLM attains the strongest clean-data score (F1 ≈ 0.933) yet is most brittle under character noise, whereas TextRCNN, despite a lower baseline, preserves comparatively higher stability in heavily perturbed conditions. Finally, we test a pragmatic hardening strategy—data augmentation with noisy variants followed by retraining—which consistently narrows robustness gaps across architectures without materially sacrificing clean-data accuracy. The benchmark and code enable reproducible evaluation and future robust modeling in cybersecurity.
Papers List
List of archived papers
ارائه مدل یادگیری ماشین برای پیشبینی سریزمانی باینری از دیدگاه مسئلههای دستهبندی با کاربرد در پیشبینی نتهای موسیقی
نیلوفر ع��دلخانی - حسام عمرانپور
بهبود هزینههای تراکنش در معماری مدیریت زنجیرهی تامین مبتنی بر زنجیرهی بلوکی
مژگان نوروزی نژاد - دکتر زهرا موحدی مژگان نوروزی نژاد - زهرا موحدی -
خوشه بندی شبکههای بیسیم ادهاک مبتنی بر محدودیتهای فازی
پروا کلیبری - کریم صمدزمینی
A novel approach audio watermarking based on (GBT,DCT,SVD)
Mahdi Mosleh
Task Scheduling for Real-time Object Detection: Methods and Performance Comparison in ADAS Applications
Mahdi Seyfipoor - Sayyed Muhammad Jaffry - Siamak Mohamadi
FiReT: A Neural Radiance Fields Framework for Wireless Field Reconstruction and Transmitter Placement
Negar Pouya - Armin Soleymani - Gholamreza Moradi - Farzaneh Abdollahi
Improved Weighting in the Automated Texts Classification using Fuzzy Method
Hamidreza Sadrarhami - S. Mohammadali Zanjani - Ghazanfar Shahgholian
Sigma: A Secure Federated Network Gaming Platform
Keyhan Mohammadi - Reza Ebrahimi Atani
پیش بینی بیماری قلبی با استفاده از روش تحلیل شبکه ای
هدیه مشتاقی محمدزاده - فاطمه باقری
From Faces to Words: An Efficient Persian Visual Lip Reading
Mana Amini - Sajjad Aemmi - Azadeh Ashouri - Reza Akhoundzadeh - Kourosh Hassanzadeh - Mohammad Reza Mohammadi
more
Samin Hamayesh - Version 42.5.2