{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2607.21155", "kind": "paper", "url": "https://arxiv.org/abs/2607.21155"}], "authors": ["Hanseok Oh", "Parishad BehnamGhader", "Benno Krojer", "Hyunji Lee", "Paul Liang", "Siva Reddy", "Verna Dankers"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2607.21155", "source_url": "https://arxiv.org/abs/2607.21155"}, "categories": ["Factuality", "General AI", "Information retrieval", "Reasoning", "Vision & 3D"], "description": {"en": "CRAG-MM-Diagnostics is a diagnostic benchmark for knowledge-intensive visual question answering (KI-VQA) with stage-wise annotations to isolate visual grounding, object identification, and knowledge retrieval/reasoning, including metadata like target ROIs and visual complexity scores."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2607.21155", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2607.21155", "title": "CRAG-MM-Diagnostics"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2607.21155", "modality": null, "name": "CRAG-MM-Diagnostics", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "High", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "unknown", "origin_record_id": "bm_crag-mm-diagnostics_8195c595", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "fad36615b56e7a3cad48931c4b0bd5d7f868975e5912080fdeb033dfef6362fb", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2607.21155"}, "publisher": null, "released": "2026-07-23", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2607.21155", "source_url": "https://arxiv.org/abs/2607.21155"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2607-21155", "source": "claire_radar", "source_benchmark_id": "2607.21155", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Vision & 3D", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2607.21155", "observedAt": "2026-10-04T03:02:42.754940Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "state": "unavailable"}}, "status": "unavailable"}, "capabilities": ["Reasoning", "Information retrieval", "Factuality"], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-20T15:22:10.044731Z", "inputHash": "992d635f361217d634d58c3a48155ba4b595996d3322d10dcf5c4d9e9cf7b0d3", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "CRAG-MM-Diagnostics is a diagnostic benchmark for knowledge-intensive visual question answering (KI-VQA) with stage-wise annotations to isolate visual grounding, object identification, and knowledge retrieval/reasoning, including metadata like target ROIs and visual complexity scores.", "domainCuration": {"method": "rules-v1", "state": "auto"}, "evidence": {"reasonCodes": ["exact coined title identity tied to benchmark evidence", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "To analyze the full KI-VQA pipeline, we introduce CRAG-MM-Diagnostics, a diagnostic benchmark with stage-wise data annotations that isolate 1) language-based visual grounding, 2) object identification, and 3) knowledge retrieval and reasoning."}, "familyId": "bmf_572d27c0d445", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_crag-mm-diagnostics_8195c595", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2607.21155", "paper": "https://arxiv.org/abs/2607.21155", "pdf": "https://arxiv.org/pdf/2607.21155", "project": null, "report": "https://arxiv.org/abs/2607.21155"}, "metrics": [], "motivation": "Knowledge-Intensive Visual Question Answering (KI-VQA) benchmarks evaluate Vision-Language Models (VLMs) as multimodal knowledge assistants by requiring external information beyond a provided image to answer questions.", "name": "CRAG-MM-Diagnostics", "oneLine": "CRAG-MM-Diagnostics is a diagnostic benchmark for knowledge-intensive visual question answering (KI-VQA) with stage-wise annotations to isolate visual grounding, object identification, and knowledge retrieval/reasoning, including metadata like target ROIs and visual complexity scores.", "paperTitle": "CRAG-MM-Diagnostics: Enabling Stage-Wise Analysis of Knowledge-Intensive VQA", "primaryDomain": "General AI", "publication": {"evidence": "Accepted to ECCV 2026", "evidenceLevel": "author-claim", "evidenceUrl": "https://arxiv.org/abs/2607.21155", "source": "arxiv-comments", "status": "acceptance_claimed", "venue": "ECCV 2026", "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "CRAG-MM-Diagnostics Team", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://arxiv.org/abs/2607.21155"}], "ranking": {"90d": {"confidence": "Low", "coverage": 0.0, "growth": {"components": {"githubStars": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfDatasetDownloads": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}, "hfPaperUpvotes": {"normalization": "percentile", "normalized": null, "percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"normalization": "log-max", "normalized": null, "value": null}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "rank": null, "score": null}, "method": "attention-ranking-v13", "rank": null, "score": null}}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-07-23", "source": {"authors": ["Hanseok Oh", "Parishad BehnamGhader", "Benno Krojer", "Hyunji Lee", "Paul Liang", "Siva Reddy", "and Verna Dankers"], "categories": ["cs.CV", "cs.AI"], "id": "2607.21155", "title": "CRAG-MM-Diagnostics: Enabling Stage-Wise Analysis of Knowledge-Intensive VQA", "type": "arxiv", "url": "https://arxiv.org/abs/2607.21155"}, "sourceUpdatedAt": "2026-07-23T10:37:01Z", "topics": ["Reasoning"], "venueAttempts": [{"decisionRaw": "Accepted to ECCV 2026", "evidence": [{"level": "author-claim", "observedAt": "2026-10-14T01:46:58.058199Z", "rawValue": "Accepted to ECCV 2026", "sourceType": "arxiv-comments", "sourceUrl": "https://arxiv.org/abs/2607.21155"}], "reviewStatus": "accepted", "venueName": "ECCV 2026"}], "whyItMatters": "End-task accuracy alone obscures failure sources in KI-VQA; this benchmark enables stage-wise analysis to identify bottlenecks, guiding improvements in multimodal retrieval-augmented generation."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
