Researchers have introduced SoccerRef-Agents, a multi-agent system designed to automate soccer refereeing with enhanced accuracy and explainability. The framework incorporates a new benchmark dataset, SoccerRefBench, featuring over 1,200 theory questions and 600 foul video clips. It also utilizes a knowledge base, RefKnowledgeDB, derived from official soccer rules and case studies, enabling precise, knowledge-driven decision-making. The system's novel multi-agent architecture employs cross-modal Retrieval-Augmented Generation (RAG) to connect visual information with regulatory texts, outperforming general-purpose multimodal large language models in evaluations. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT This research could lead to more objective and consistent officiating in sports through advanced AI.
RANK_REASON This is a research paper introducing a novel multi-agent system and benchmark for automated soccer refereeing.