{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2606.04773", "kind": "paper", "url": "https://arxiv.org/abs/2606.04773"}], "authors": ["Yong Cao", "Chuqiao Li", "Xianghui Xie", "Gerard Pons-Moll", "Andreas Geiger"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.04773", "source_url": "https://arxiv.org/abs/2606.04773"}, "categories": ["Multimodal", "Robotics & Embodied AI"], "description": {"en": "NextMotionQA is a benchmark for human motion understanding with VLMs, featuring three tasks: multiple-choice QA, video captioning, and fine-grained error correction. It includes expert-verified annotations across three semantic axes and three complexity levels."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.04773", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.04773", "title": "NextMotionQA"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Tübingen AI Center", "University of Tübingen", "Max Planck Institute for Informatics"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.04773", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.04773", "source_url": "https://openalex.org/W7163525931"}, "key": "claire-radar:2606.04773", "modality": null, "name": "NextMotionQA", "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_nextmotionqa_b6672a26", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "58d373ba32c7b5fe684b1717e53f0b5c0a2191d576cbdb8ef681679609b3460a", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2606.04773"}, "publisher": null, "released": "2026-06-03", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.04773", "source_url": "https://arxiv.org/abs/2606.04773"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-04773", "source": "claire_radar", "source_benchmark_id": "2606.04773", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["Robotics & Embodied AI"], "area": "Multimodal", "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/2606.04773", "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": [], "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": "e1f87425cff169224320dfaab39b5fd3689c3313d1d31a9b59365f5ca7adce53", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "NextMotionQA is a benchmark for human motion understanding with VLMs, featuring three tasks: multiple-choice QA, video captioning, and fine-grained error correction. It includes expert-verified annotations across three semantic axes and three complexity levels.", "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 bridge this gap, we introduce NextMotionQA, a comprehensive benchmark that leverages vision-language models (VLMs) for semi-automated, expert-verified dataset."}, "familyId": "bmf_cfb80271e1dd", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_nextmotionqa_b6672a26", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": ["Robotics"], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2606.04773", "paper": "https://arxiv.org/abs/2606.04773", "pdf": "https://arxiv.org/pdf/2606.04773", "project": null, "report": "https://arxiv.org/abs/2606.04773"}, "metrics": [], "motivation": "Reliable evaluation of human motion understanding is fundamental to advancing embodied AI, robotics, and animation.", "name": "NextMotionQA", "oneLine": "NextMotionQA is a benchmark for human motion understanding with VLMs, featuring three tasks: multiple-choice QA, video captioning, and fine-grained error correction. It includes expert-verified annotations across three semantic axes and three complexity levels.", "paperTitle": "NextMotionQA: Benchmarking and Judging Human Motion Understanding with Vision-Language Models", "primaryDomain": "Robotics & Embodied AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.04773", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "publishers": [{"name": "arXiv", "organizationType": "benchmark-organization", "role": "benchmark-publisher", "sourceUrl": "https://arxiv.org/abs/2606.04773"}], "ranking": {}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-06-03", "source": {"authors": ["Yong Cao", "Chuqiao Li", "Xianghui Xie", "Gerard Pons-Moll", "Andreas Geiger"], "categories": ["cs.CV", "cs.CL"], "id": "2606.04773", "title": "NextMotionQA: Benchmarking and Judging Human Motion Understanding with Vision-Language Models", "type": "arxiv", "url": "https://arxiv.org/abs/2606.04773"}, "sourceUpdatedAt": "2026-06-03T11:53:57Z", "topics": ["Multimodal"], "venueAttempts": [], "whyItMatters": "Existing motion benchmarks have coarse granularity and ambiguity. NextMotionQA provides structured tasks across complexity levels, enabling diagnosis of VLM capability gaps in motion understanding and judging."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
