SPRINT: Semi-supervised Prototypical Representation for Few-Shot Class-Incremental Tabular Learning
This work provides a novel FSCIL framework specifically for tabular data, which is crucial for real-world systems in domains like cybersecurity and healthcare that need to adapt to new concepts with limited labeled data.
The paper addresses few-shot class-incremental learning (FSCIL) for tabular data, a domain largely unexplored compared to computer vision. The authors introduce SPRINT, a framework that leverages abundant unlabeled data and low storage costs in tabular streams through a mixed episodic training strategy and confidence-based pseudo-labeling. SPRINT achieves a state-of-the-art average accuracy of 77.37% (5-shot) across six diverse benchmarks, outperforming the strongest incremental baseline by 4.45%.
Real-world systems must continuously adapt to novel concepts from limited data without forgetting previously acquired knowledge. While Few-Shot Class-Incremental Learning (FSCIL) is established in computer vision, its application to tabular domains remains largely unexplored. Unlike images, tabular streams (e.g., logs, sensors) offer abundant unlabeled data, a scarcity of expert annotations and negligible storage costs, features ignored by existing vision-based methods that rely on restrictive buffers. We introduce SPRINT, the first FSCIL framework tailored for tabular distributions. SPRINT introduces a mixed episodic training strategy that leverages confidence-based pseudo-labeling to enrich novel class representations and exploits low storage costs to retain base class history. Extensive evaluation across six diverse benchmarks spanning cybersecurity, healthcare, and ecological domains, demonstrates SPRINT's cross-domain robustness. It achieves a state-of-the-art average accuracy of 77.37% (5-shot), outperforming the strongest incremental baseline by 4.45%.