TIGER: Mitigating Hallucinations in Multimodal Generation
TIGER utilizes evidence routing graphs to detect and repair factual errors in AI-generated content from images, audio, and video.
Sources arxiv.org
TIGER utilizes evidence routing graphs to detect and repair factual errors in AI-generated content from images, audio, and video.
DeepSciVerify employs a two-stage process to verify the alignment between scientific claims and cited sources, enhancing the reliability of LLM-generated reports.