This paper introduces a framework for auditable decision management in telecom and IoT fraud control, utilizing blockchain technology. The research compares different machine learning approaches, finding that a QLoRA-tuned LLM branch is more usable than zero-shot prompting but does not outperform a lower-cost centralized ensemble. While the LLM approach shows promise in usability, evaluations on synthetic data and a replay corpus indicate that a centralized ML model (M1) offers a better balance of performance metrics. AI
IMPACT This research could lead to more robust and auditable fraud detection systems in telecom and IoT, potentially improving accuracy and transparency.
RANK_REASON The cluster contains an academic paper detailing a novel framework and evaluation of machine learning models for a specific application.
AI-generated summary · Google Gemini · from 2 sources. How we write summaries →
