Foundation Models for Cybersecurity: A Comprehensive Multi-Modal Evaluation of TabPFN and TabICL for Tabular Intrusion Detection

Abstract

Este estudio presenta una evaluación exhaustiva de modelos fundacionales aplicados a la detección de intrusiones en datos tabulares, comparando TabPFN, TabICL y modelos de lenguaje con métodos tradicionales como Random Forest y TabNet. Utilizando tres conjuntos de datos de ciberseguridad (CIC-IDS2017, N-BaIoT y CIC-UNSW), se demuestra que los modelos fundacionales superan a los enfoques clásicos, especialmente en la detección de clases minoritarias como Heartbleed e Infiltration. TabICL alcanza una precisión del 99.59% sin necesidad de ajuste de hiperparámetros ni reentrenamiento, lo que representa un cambio de paradigma hacia arquitecturas adaptativas y ligeras. El estudio propone una arquitectura de detección en tres niveles para centros de operaciones de seguridad, combinando eficiencia computacional con cobertura integral de amenazas. Se concluye que los modelos fundacionales ofrecen ventajas significativas para sistemas de ciberseguridad de próxima generación.While traditional ensemble methods have dominated tabular intrusion detection systems (IDSs), recent advances in foundation models present new opportunities for enhanced cybersecurity applications. This paper presents a comprehensive multi-modal evaluation of foundation models—specifically TabPFN (Tabular Prior-Data Fitted Network), TabICL (Tabular In-Context Learning), and large language models—against traditional machine learning approaches across three cybersecurity datasets: CIC-IDS2017, N-BaIoT, and CIC- UNSW. Our rigorous experimental framework addresses critical methodological challenges through model-appropriate evaluation protocols and comprehensive assessment across multiple data variants. Results demonstrate that foundation models achieve superior and more consistent performance compared with traditional approaches, with TabPFN and TabICL establishing new state-of-the-art results across all datasets. Most significantly, these models uniquely achieve non-zero recall across all classes, including rare threats like Heartbleed and Infiltration, while traditional ensemble methods—despite achieving >99% overall accuracy—completely fail on several minority classes. TabICL demonstrates partic- ularly strong performance on CIC-IDS2017 (99.59% accuracy), while TabPFN maintains consistent performance across all datasets, suggesting robust generalization capabilities. Both foundation models achieve these results using only fractions of the available train- ing data and requiring no hyperparameter tuning, representing a paradigm shift toward training-light, hyperparameter-free adaptive IDS architectures, where TabPFN requires no task-specific fitting and TabICL leverages efficient in-context adaptation without retraining. Cross-dataset validation reveals that foundation models maintain performance advantages across diverse threat landscapes, while traditional methods exhibit significant dataset- specific variations. These findings challenge the cybersecurity community’s reliance on tree-based ensembles and demonstrate that foundation models offer superior capabilities for next-generation intrusion detection systems in IoT environments