AI Paper Atlas

Research map / Vision-language and multimodal / Visual question answering

CRAG-MM-Diagnostics: Enabling Stage-Wise Analysis of Knowledge-Intensive VQA

Hanseok Oh, Parishad BehnamGhader, Benno Krojer, Hyunji Lee, Paul Pu Liang, Siva Reddy, Verna Dankers

ECCV 2026 · Vision-language and multimodal › Visual question answering · approach: Universal VQA models · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add disambiguated queries for ambiguous referring expressions (ablation)visual grounding VLMs · CRAG-MM-Diagnostics groundinggrounding accuracy: mean 44.0% improvementImproved
add disambiguated queries (ablation)Llama and GPT-5-mini grounding · CRAG-MM-Diagnostics groundinggrounding accuracyNo clear effect
crop the target region instead of full image (ablation)generalized VLMs · Object identificationaccuracyHurt
keep original question instead of simplified prompt (ablation)generalized VLMs · Object identificationaccuracy: 7.9 percentage pointsImproved
crop the target region before image retrieval (ablation)specialized image retrievers · Object identification recall@10recall@10Improved
add the question text to multimodal retrievers (ablation)VLM2Vec2 and Qwen-3-VL-Embedding · Object identification recall@10recall@10No clear effect
provide gold entity name (text-only question) (ablation)generalized VLMs · KI-VQA accuracyaccuracy: 8.7 points on averageImproved
add disambiguated query (ablation)generalized VLMs · KI-VQA accuracy (ambiguous)accuracy: 1.8-point gainImproved
add image retrieval (ablation)Qwen and GPT-5 · KI-VQA accuracyaccuracy: 6.8-point average boostImproved
add text retrieval (ablation)VLMs · KI-VQA accuracyaccuracy: 6.3 pointsImproved
use entity-name text-only questions for text retrieval (ablation)GPT-5 and Qwen · KI-VQA accuracyaccuracy: 16.6 and 26.2 pointsImproved
add visual grounding crop before image retrieval (ablation)bimodal RAG · KI-VQA accuracyaccuracy: 13.3 and 8.5 percentage pointsImproved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy