TeleAntiFraud-28k 📞🛡️

🎉 News

• [2026.04] We launched the TeleAntiFraud open-source community for telecom anti-fraud research: https://teleantifraud.github.io/ 🌐
• [2026.04] SAFE-QAQ: End-to-End Slow-Thinking Audio-Text Fraud Detection via Reinforcement Learning has been accepted by ACL 2026! 🎊
• [2025] TeleAntiFraud-28k: An Audio-Text Slow-Thinking Dataset for Telecom Fraud Detection has been accepted by ACM MM 2025! 🎊
• [2026.04] A sanitized public release with dataset card, audio archives, and preview samples is now available on Hugging Face and ModelScope. 📚
• [Dataset] TeleAntiFraud-28k is available on ModelScope. 📚

TeleAntiFraud-28k is the first open-source audio-text slow-thinking dataset specifically designed for automated telecom fraud analysis. This dataset integrates audio signals with reasoning-oriented textual analysis, providing high-quality multimodal training data for telecom fraud detection research. 🔍💡

📊 Dataset Overview

• Total Samples: 28,511 rigorously processed speech-text pairs 📋
• Total Audio Duration: 307 hours ⏱️
• Unique Feature: Detailed annotations for fraud reasoning 🧠
• Task Categories: Scenario classification, fraud detection, fraud type classification 🎯

Public Dataset Release

We provide a sanitized public release for direct download and benchmarking:

• Hugging Face dataset page
• ModelScope dataset page
• packaged binary_classification.zip, sft.zip, and audio.zip
• preview audio samples directly playable on the dataset page
• normalized relative audio paths without machine-specific absolute paths

🏗️ Dataset Construction Strategies

1. 🔒 Privacy-preserved Text-Truth Sample Generation

• Using ASR-transcribed call recordings (with anonymized original audio)
• Ensuring real-world consistency through TTS model regeneration
• Strict adherence to privacy protection standards

2. 🚀 Semantic Enhancement

• LLM-based self-instruction sampling on authentic ASR outputs
• Expanding scenario coverage to improve model generalization
• Enriching the diversity of conversational contexts

3. 🤖 Multi-agent Adversarial Synthesis

• Simulation of emerging fraud tactics
• Generation through predefined communication scenarios and fraud typologies
• Enhancing dataset adaptability to new fraud techniques

🎯 TeleAntiFraud-Bench

We have constructed TeleAntiFraud-Bench, a standardized evaluation benchmark comprising proportionally sampled instances from TeleAntiFraud-28k, to facilitate systematic testing of model performance and reasoning capabilities on telecom fraud detection tasks. 📐✅

Evaluation Utilities

We provide sanitized evaluation scripts in evaluation/, including:

• classification metrics for scenario classification, fraud detection, and fraud type classification
• preparation of model outputs for reasoning assessment
• an OpenAI-compatible LM-as-judge runner
• the probability-based reasoning-quality judging prompt

The ACL 2026 paper repository SAFE-QAQ, which reports results on TeleAntiFraud, is available at Control-derek/SAFE-QAQ.

🤖 Model Contribution

We contribute a production-optimized supervised fine-tuning (SFT) model based on Qwen2-Audio, trained on the TeleAntiFraud training set. 🎨⚡

Released Model

• Hugging Face model: JimmyMa99/AntiFraud-SFT
• ModelScope model: JimmyMa99/AntiFraud-SFT
• base model: Qwen/Qwen2-Audio-7B-Instruct
• model type: supervised fine-tuning for audio-text telecom fraud detection

📝 Examples

Explore our dataset examples to better understand the telecom fraud detection capabilities: 👀

• Case 1: Normal Conversation Analysis - Detailed analysis of a legitimate phone conversation ✅
• Case 2: Fraud Conversation Analysis - Step-by-step reasoning for detecting a fraudulent call ⚠️
• Evaluation Sample - Representative sample from our evaluation benchmark 📊
• Model Output: Normal Conversation - Our model's reasoning process on a legitimate call 🤖✅
• Model Output: Fraud Detection - Model's analysis and detection of a fraudulent call 🤖⚠️

🛠️ Multi-Agent Data Collection

To collect fraudulent conversation data: 💼

1. Insert your API key in multi-agents-tools/AntiFraudMatrix/main.py (uses SiliconFlow API key) 🔑
2. Run the following command to generate fraudulent dialog text:

   python multi-agents-tools/AntiFraudMatrix/main.py

3. Results will be saved in the result directory 📁

For normal conversation data: 💬

• Use multi-agents-tools/AntiFraudMatrix-normal/main.py following the same process

🎙️ Voice Synthesis with ChatTTS

To synthesize speech from the collected text: 🔊

1. Install the necessary dependencies 📦

2. Run the API server:

   fastapi dev ChatTTS/examples/api/main_new_new.py --host 0.0.0.0 --port 8006

3. Use any of the scripts in ChatTTS/examples/api/normal_run*.sh or ChatTTS/examples/api/run*.sh 🚀

   Modify the port in these scripts if needed, then run:

   bash ChatTTS/examples/api/run*.sh

🌟 Open-Source Resources

• TeleAntiFraud public dataset release (Hugging Face) 🤗
• TeleAntiFraud public dataset release (ModelScope) 📚
• AntiFraud-SFT model (Hugging Face) 🤖
• AntiFraud-SFT model (ModelScope) 🤖
• SAFE-QAQ code repository 🧪
• TeleAntiFraud-28k dataset 📚
• TeleAntiFraud-Bench evaluation benchmark 🏆
• Evaluation and LM-as-judge utilities ⚖️
• Data processing framework (supporting community-driven dataset expansion) 🔧
• TeleAntiFraud-Qwen2-Audio SFT model 🤖

🎯 Key Contributions

1. Establishing a foundational framework for multimodal anti-fraud research 🏗️
2. Addressing critical challenges in data privacy and scenario diversity 🔐
3. Providing high-quality training data for telecom fraud detection 📈
4. Open-sourcing data processing tools to enable community collaboration 🤝

🙏 Acknowledgements

We would like to express our sincere gratitude to all the organizations and individuals who have provided invaluable support throughout this project: ❤️

• China Mobile Internet Company (中移互联网) - For their industry expertise and technical guidance 🏢
• Intern Community (书生社区) - For their open-source ecosystem support and collaboration 🌍
• ModelScope Community (魔搭社区) - For their platform support and community resources 🎪
• SmartFlowAI Community (机智流社区) - For their technical contributions and collaborative efforts 💡
• Control-derek - For his technical expertise and valuable contributions 👨‍💻
• vansin - For his dedicated support and assistance 🤝
• Jintao-Huang - For his valuable suggestions and contributions 💭

Their contributions have been instrumental in making this project a success and advancing the field of telecom fraud detection research. 🚀

📄 Citation

@inproceedings{ma2025teleantifraud,
  title={TeleAntiFraud-28k: An Audio-Text Slow-Thinking Dataset for Telecom Fraud Detection},
  author={Ma, Zhiming and Wang, Peidong and Huang, Minhua and Wang, Jinpeng and Wu, Kai and Lv, Xiangzhao and Pang, Yachun and Yang, Yin and Tang, Wenjie and Kang, Yuchen},
  booktitle={Proceedings of the 33rd ACM International Conference on Multimedia},
  pages={5853--5862},
  year={2025}
}

@article{wang2026safe,
  title={SAFE-QAQ: End-to-End Slow-Thinking Audio-Text Fraud Detection via Reinforcement Learning},
  author={Wang, Peidong and Ma, Zhiming and Dai, Xin and Liu, Yongkang and Feng, Shi and Yang, Xiaocui and Hu, Wenxing and Wang, Zhihao and Pan, Mingjun and Yuan, Li and others},
  journal={arXiv preprint arXiv:2601.01392},
  year={2026}
}