Reasoner for Real-World Event Detection: Scaling Reinforcement Learning via Adaptive Perplexity-Aware Sampling Strategy

TL;DR

This paper introduces APARL, a reinforcement learning framework that uses adaptive, perplexity-aware sampling and curriculum learning to improve abnormal event detection in customer service dialogues, especially in out-of-domain scenarios.

Contribution

The paper presents a novel dual-loop curriculum learning architecture that enhances the reasoning and out-of-domain generalization of large language models for event detection.

Findings

Achieves a 17.19% improvement in F1 score on food delivery dialogue tasks.

Enhances out-of-domain transferability by 9.59%.

Demonstrates robustness and adaptability in real-world industrial scenarios.

Abstract

Detecting abnormal events in real-world customer service dialogues is highly challenging due to the complexity of business data and the dynamic nature of customer interactions. Moreover, models must demonstrate strong out-of-domain (OOD) generalization to enable rapid adaptation across different business scenarios and maximize commercial value. In this work, we propose a novel Adaptive Perplexity-Aware Reinforcement Learning (APARL) framework that leverages the advanced reasoning capabilities of large language models for abnormal event detection. APARL introduces a dual-loop dynamic curriculum learning architecture, enabling the model to progressively focus on more challenging samples as its proficiency increases. This design effectively addresses performance bottlenecks and significantly enhances OOD transferability. Extensive evaluations on food delivery dialogue tasks show that our model achieves significantly enhanced adaptability and robustness, attaining the highest F1 score with an average improvement of 17.19\%, and an average improvement of 9.59\% in OOD transfer tests. This method provides a superior solution for industrial deployment of anomaly detection models, contributing to improved operational efficiency and commercial benefits.