用强化学习实现端到端语音反欺诈，不依赖转录文本。
SAFE-QAQ: End-to-End Slow-Thinking Audio-Text Fraud Detection via Reinforcement Learning
- 直接分析原始语音，避开语音识别错误。
- 通过分层推理机制，准确捕捉声调等细微线索。
- 支持实时风险评估，适合金融客服场景使用。
现有欺诈检测方法主要依赖语音转文字，受语音识别误差影响，且遗漏声调、环境背景等关键音频特征，难以应对复杂欺骗行为。为此，我们提出端到端的语音慢思考反欺诈框架 SAFE-QAQ。该框架消除转录错误对检测性能的影响；设计基于规则的慢思考奖励机制，通过分层推理过程精准捕捉细粒度音频特征，系统识别欺诈模式；同时引入动态风险评估机制，支持通话中的实时欺诈预警。在 TeleAntiFraud-Bench 数据集上的实验表明，SAFE-QAQ 在准确率、推理效率和实时处理能力上均显著优于现有方法。目前该系统已部署，每日分析超 7 万通电话，有效自动化复杂欺诈检测，大幅降低人工成本与财务损失。
原文摘要 · Abstract (English)
Existing fraud detection methods predominantly rely on transcribed text, suffering from ASR errors and missing crucial acoustic cues like vocal tone and environmental context. This limits their effectiveness against complex deceptive strategies. To address these challenges, we first propose \textbf{SAFE-QAQ}, an end-to-end comprehensive framework for audio-based slow-thinking fraud detection. First, the SAFE-QAQ framework eliminates the impact of transcription errors on detection performance. Secondly, we propose rule-based slow-thinking reward mechanisms that systematically guide the system to identify fraud-indicative patterns by accurately capturing fine-grained audio details, through hierarchical reasoning processes. Besides, our framework introduces a dynamic risk assessment framework during live calls, enabling early detection and prevention of fraud. Experiments on the TeleAntiFraud-Bench demonstrate that SAFE-QAQ achieves dramatic improvements over existing methods in multiple key dimensions, including accuracy, inference efficiency, and real-time processing capabilities. Currently deployed and analyzing over 70,000 calls daily, SAFE-QAQ effectively automates complex fraud detection, reducing human workload and financial losses. Code: https://anonymous.4open.science/r/SAFE-QAQ.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。