Saved in:
Bibliographic Details
Main Authors: Singla, Saurav, Singla, Aarav, Gupta, Advik, Gupta, Parnika
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.20019
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909708403605504
author Singla, Saurav
Singla, Aarav
Gupta, Advik
Gupta, Parnika
author_facet Singla, Saurav
Singla, Aarav
Gupta, Advik
Gupta, Parnika
contents We propose a meta learning framework for detecting anomalies in human language across diverse domains with limited labeled data. Anomalies in language ranging from spam and fake news to hate speech pose a major challenge due to their sparsity and variability. We treat anomaly detection as a few shot binary classification problem and leverage meta-learning to train models that generalize across tasks. Using datasets from domains such as SMS spam, COVID-19 fake news, and hate speech, we evaluate model generalization on unseen tasks with minimal labeled anomalies. Our method combines episodic training with prototypical networks and domain resampling to adapt quickly to new anomaly detection tasks. Empirical results show that our method outperforms strong baselines in F1 and AUC scores. We also release the code and benchmarks to facilitate further research in few-shot text anomaly detection.
format Preprint
id arxiv_https___arxiv_org_abs_2507_20019
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Anomaly Detection in Human Language via Meta-Learning: A Few-Shot Approach
Singla, Saurav
Singla, Aarav
Gupta, Advik
Gupta, Parnika
Computation and Language
Artificial Intelligence
We propose a meta learning framework for detecting anomalies in human language across diverse domains with limited labeled data. Anomalies in language ranging from spam and fake news to hate speech pose a major challenge due to their sparsity and variability. We treat anomaly detection as a few shot binary classification problem and leverage meta-learning to train models that generalize across tasks. Using datasets from domains such as SMS spam, COVID-19 fake news, and hate speech, we evaluate model generalization on unseen tasks with minimal labeled anomalies. Our method combines episodic training with prototypical networks and domain resampling to adapt quickly to new anomaly detection tasks. Empirical results show that our method outperforms strong baselines in F1 and AUC scores. We also release the code and benchmarks to facilitate further research in few-shot text anomaly detection.
title Anomaly Detection in Human Language via Meta-Learning: A Few-Shot Approach
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2507.20019