Research map / Graph learning / Graph self-supervised learning
TriFusion-IDS: A Multimodal Graph-Tabular-Text Contrastive Framework for Cross-Dataset Intrusion Detection
Qinxin Zhao, Sheng Zhong
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add graph module (ablation) | TriFusion-IDS · CSE-CIC-IDS -> ToN-IoT | Accuracy (%): 55.17 -> 66.65 | Improved |
| add tabular module (ablation) | TriFusion-IDS · CSE-CIC-IDS -> UNSW-NB15 | Macro F1 (%): 53.25 -> 64.68 | Improved |
| add graph perturbation augmentation (ablation) | TriFusion-IDS · CSE-CIC-IDS -> ToN-IoT | Accuracy (%): 48.85 -> 66.65 | Improved |
| add text prompting (ablation) | TriFusion-IDS · CSE-CIC-IDS -> UNSW-NB15 | Macro F1 (%): 51.03 -> 64.68 | Improved |
| use anomaly-center inference instead of multi-class similarity (ablation) | TriFusion-IDS · CSE-CIC-IDS -> ToN-IoT | Accuracy (%): 46.66 -> 66.65 | Improved |
| use anomaly-center inference instead of single anomaly prompt (ablation) | TriFusion-IDS · CSE-CIC-IDS -> ToN-IoT | Accuracy (%): 54.26 -> 66.65 | Improved |
| perturbation loss weight lambda raised from 0.1 to 0.5, then beyond (ablation) | TriFusion-IDS · cross-dataset Macro F1 | Macro F1 (%) | Mixed |
| compare with GraphCLIP (adds tabular features and anomaly center) | TriFusion-IDS · UNSW-NB15 -> ToN-IoT (100%) | Macro F1 (%): 51.61 -> 57.93 | Improved |
| TriFusion-IDS vs TabNet on CSE-CIC-IDS -> UNSW-NB15 (100%) | TriFusion-IDS · CSE-CIC-IDS -> UNSW-NB15 (100%) | Accuracy (%): 98.72 (TabNet) -> 93.80 | Hurt |
| use 10% instead of 100% of labeled source data | TriFusion-IDS · UNSW-NB15 -> CSE-CIC-IDS | Accuracy (%): 96.75 -> 95.72 | No clear effect |
Closest papers by meaning
- Self-Supervised Learning of Graph Representations for Network Intrusion Detection (NeurIPS 2025 · similarity 0.84)
- DiffGAD: A Diffusion-based Unsupervised Graph Anomaly Detector (ICLR 2025 · similarity 0.82)
- UniFORM: Towards Unified Framework for Anomaly Detection on Graphs (AAAI 2025 · similarity 0.82)
- Federated Graph Anomaly Detection Through Contrastive Learning with Global Negative Pairs (AAAI 2025 · similarity 0.81)
- DiffuPac: Contextual Mimicry in Adversarial Packets Generation via Diffusion Model (NeurIPS 2024 · similarity 0.81)
- Revisiting Graph Contrastive Learning on Anomaly Detection: A Structural Imbalance Perspective (AAAI 2025 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
