Multimedia Verification Through Multi-Agent Deep Research Multimodal Large Language Models

2025-07-06

AI Summary

To address the challenges of geolocation, temporal attribution, and cross-platform provenance in multimedia misinformation detection, this paper proposes a multi-agent collaborative automated verification framework. Centered on multimodal large language models (MLLMs), the framework integrates reverse image search, metadata parsing, fact-checking databases, and authoritative news source processing into a six-stage deep analytical pipeline, enabling context-aware cross-verification across spatial, temporal, provenance, and motivational dimensions. Its key innovation lies in the first decoupled yet coordinated integration of MLLM-driven deep-reasoning agents with domain-specific toolchains, supporting dynamic task dispatching and evidence-based closed-loop validation. Evaluated on multiple challenging benchmarks, the system achieves substantial improvements: +23.6% in geolocation accuracy, +18.4% in temporal inference accuracy, and +31.2% in cross-platform provenance success rate—thereby enabling explainable, trustworthy authenticity assessment in complex misinformation scenarios.

Technology Category

• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)
• Machine Learning: Large Multimodal Models (LMMs)
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Web Mining and Content Analysis: Mining multimedia, multimodal, multilingual, cross-lingual Web data
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Responsible Web: Mitigating misinformation and disinformation

Abstract

This paper presents our submission to the ACMMM25 - Grand Challenge on Multimedia Verification. We developed a multi-agent verification system that combines Multimodal Large Language Models (MLLMs) with specialized verification tools to detect multimedia misinformation. Our system operates through six stages: raw data processing, planning, information extraction, deep research, evidence collection, and report generation. The core Deep Researcher Agent employs four tools: reverse image search, metadata analysis, fact-checking databases, and verified news processing that extracts spatial, temporal, attribution, and motivational context. We demonstrate our approach on a challenge dataset sample involving complex multimedia content. Our system successfully verified content authenticity, extracted precise geolocation and timing information, and traced source attribution across multiple platforms, effectively addressing real-world multimedia verification scenarios.

Problem

• Detect multimedia misinformation using MLLMs and verification tools
• Verify content authenticity and extract geolocation information
• Trace source attribution across multiple platforms effectively

Innovation

• Multi-agent system with MLLMs and verification tools
• Six-stage process including deep research and evidence collection
• Four specialized tools for comprehensive multimedia analysis