mlsanityData to 14 Oct 2026

SMETA-ZSL:Semantic Meta-Alignment for Zero-Shot Threat Classification

arXiv 2607.09936 ↗First posted 10 Jul 2026cs.LG cs.AI cs.CRcited by 0

Not on the map. Why?

Abstract (arXiv)

Cybersecurity systems must adapt rapidly to emerging threats. However, labeled data for new threat categories is unavailable when those threats first appear. Generalized zero-shot learning offers a natural solution by enabling recognition of unseen classes through auxiliary semantic knowledge rather than labeled examples. Large language models are particularly promising in this setting because they can convert unstructured CTI reports into semantic prototypes for emerging threats. However, applying language-driven zero-shot learning to cybersecurity is difficult due to strong semantic overlap between threat descriptions, heterogeneity between behavioral attributes and text, severe class imbalance, and open-set conditions where unseen threats are unknown during training. We propose SMETA-ZSL, that learns semantic prototypes from overlapping language descriptions through contrastive finetuning, aligns behavioral features through episodic meta-learning and knowledge distillation, and performs adaptive routing for generalization across seen-unseen classes. Across 7 benchmarks, SMETA-ZSL delivers the strongest overall generalized zero-shot performance under the strictest inductive setting, surpassing prior methods by 10.8 points on average, with gains up to 18.1 points. Github:https://github.com/Security-And-Intelligence-Lab-UTEP/SMETA-ZSL

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 29 of its 54 references that are in the map. First posted on arXiv 10 Jul 2026.

Builds on

29 of 54 references.

References · 29

  1. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,166
  2. 2020Supervised Contrastive Learningcited by 2,829
  3. 2022TabPFN: A Transformer That Solves Small Tabular Classification Problems in a Secondcited by 674
  4. 2017Zero-Shot Learning -- The Good, the Bad and the Uglycited by 255
  5. 2020A Review of Generalized Zero-Shot Learning Methodscited by 159
  6. 2019Fine-Grained Spoiler Detection from Large-Scale Review Corporacited by 77
  7. 2017A Generative Model For Zero Shot Learning Using Conditional Variational Autoencoderscited by 64
  8. 2019r/Fakeddit: A New Multimodal Benchmark Dataset for Fine-grained Fake News Detectioncited by 55
  9. 2020Meta-KD: A Meta Knowledge Distillation Framework for Language Model Compression across Domainscited by 33
  10. 2022DUET: Cross-modal Semantic Grounding for Contrastive Zero-shot Learningcited by 33
  11. 2021Generating Fake Cyber Threat Intelligence Using Transformer-Based Modelscited by 29
  12. 2021Dual Progressive Prototype Network for Generalized Zero-Shot Learningcited by 26
17 more
  1. 2024LLMCloudHunter: Harnessing LLMs for Automated Extraction of Detection Rules from Cloud-Based CTIcited by 18
  2. 2025LAMD: Context-driven Android Malware Detection and Classification with LLMscited by 18
  3. 2024Revisiting Concept Drift in Windows Malware Detection: Adaptation to Real Drifted Malware with Minimal Samplescited by 11
  4. 2024Tabular Transfer Learning via Prompting LLMscited by 10
  5. 2024CLIP-Decoder : ZeroShot Multilabel Classification using Multimodal CLIP Aligned Representationcited by 9
  6. 2024Data-Free Generalized Zero-Shot Learningcited by 8
  7. 2025Zero-shot Meta-learning for Tabular Prediction Tasks with Adversarially Pre-trained Transformercited by 8
  8. 2024Seeing Beyond Classes: Zero-Shot Grounded Situation Recognition via Language Explainercited by 7
  9. 2022A Gating Model for Bias Calibration in Generalized Zero-shot Learningcited by 6
  10. 2022Semantically Grounded Visual Embeddings for Zero-Shot Learningcited by 4
  11. 2024Dual Expert Distillation Network for Generalized Zero-Shot Learningcited by 4
  12. 2025Evaluating LLM Generated Detection Rules in Cybersecuritycited by 4
  13. 2026CTI-REALM: Benchmark to Evaluate Agent Performance on Security Detection Rule Generation Capabilitiescited by 4
  14. 2025Semantic-Aware Contrastive Fine-Tuning: Boosting Multimodal Malware Classification with Discriminative Embeddingscited by 3
  15. 2025LLM Empowered Prototype Learning for Zero and Few-Shot Tasks on Tabular Datacited by 2
  16. 2026LLM-FS: Zero-Shot Feature Selection for Effective and Interpretable Malware Detectioncited by 1
  17. 2026CLIP-driven Zero-shot Learning with Ambiguous Labelscited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025From Retrieval to Reasoning: A Framework for Cyber Threat Intelligence NER with Explicit and Adaptive Instructions
  2. 2026LLMs for Zero-Shot Threat Detection via Structured Risk Indicators
  3. 2025Fine-tuning of Large Language Models for Domain-Specific Cybersecurity Knowledge
  4. 2025Less Data, More Security: Advancing Cybersecurity LLMs Specialization via Resource-Efficient Domain-Adaptive Continuous Pre-training with Minimal Tokens
  5. 2025Toward Cybersecurity-Expert Small Language Models
  6. 2025Adapting Large Language Models to Emerging Cybersecurity using Retrieval Augmented Generation
  7. 2024Crimson: Empowering Strategic Reasoning in Cybersecurity through Large Language Models
  8. 2025Cyber-Zero: Training Cybersecurity Agents without Runtime