{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2605.22100", "kind": "paper", "url": "https://arxiv.org/abs/2605.22100"}], "authors": ["Bangbang Zhou", "Hangdi Xing", "Yifan Chen", "Jianjun Xu", "Qi Zheng", "Feiyu Gao", "Zhibo Yang", "Shuai Bai", "Ming Yan", "Jieping Ye", "Hongtao Xie"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2605.22100", "source_url": "https://arxiv.org/abs/2605.22100"}, "categories": ["General AI", "Language & Knowledge", "cs.AI"], "description": {"en": "MPDocBench-Parse is a benchmark for multi-page document parsing, covering 433 manually annotated documents with 3,246 pages across 15 document types in English and Chinese. It evaluates content fidelity and logical structure, including text, table, formula recognition, reading order, and heading hierarchy recovery."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2605.22100", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2605.22100", "title": "MPDocBench-Parse"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2605.22100", "modality": null, "name": "MPDocBench-Parse", "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_mpdocbench-parse_f77fb798", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "08023a7d6ff8cf8326f615ba042ef822d4aaab550f61a4b1bc44f8b7a6fef2e9", "relation": "introduces", "review_state": "unreviewed", "source_url": "https://arxiv.org/abs/2605.22100"}, "publisher": null, "released": "2026-05-21", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2605.22100", "source_url": "https://arxiv.org/abs/2605.22100"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2605-22100", "source": "claire_radar", "source_benchmark_id": "2605.22100", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "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": 0, "hfPaperUrl": "https://huggingface.co/papers/2605.22100", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}}, "status": "fresh"}, "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": "71f565417af5043db56aedc917cbb88c3b877c75d07086ea7814f969a8b01ce2", "model": "deepseek-v4-flash"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "MPDocBench-Parse is a benchmark for multi-page document parsing, covering 433 manually annotated documents with 3,246 pages across 15 document types in English and Chinese. It evaluates content fidelity and logical structure, including text, table, formula recognition, reading order, and heading hierarchy recovery.", "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 address these gaps, we propose MPDocBench-Parse, a benchmark for multi-page document parsing in real-world applications."}, "familyId": "bmf_790789c36119", "firstSeenAt": "2026-08-19", "githubIndex": {"candidateCount": 0, "indexedAt": "2026-09-17T04:40:34.111665+00:00", "repositories": [], "status": "not-found"}, "heat": null, "id": "bm_mpdocbench-parse_f77fb798", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "hfPaper": "https://huggingface.co/papers/2605.22100", "paper": "https://arxiv.org/abs/2605.22100", "pdf": "https://arxiv.org/pdf/2605.22100", "project": null, "report": "https://arxiv.org/abs/2605.22100"}, "metrics": [], "motivation": "Document parsing converts visually rich documents into machine-readable structured representations, forming a crucial foundation for information systems.", "name": "MPDocBench-Parse", "oneLine": "MPDocBench-Parse is a benchmark for multi-page document parsing, covering 433 manually annotated documents with 3,246 pages across 15 document types in English and Chinese. It evaluates content fidelity and logical structure, including text, table, formula recognition, reading order, and heading hierarchy recovery.", "paperTitle": "MPDocBench-Parse: Benchmarking Practical Multi-page Document Parsing", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2605.22100", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-10-14T01:46:58.058199Z"}, "publications": [], "ranking": {}, "readiness": "Paper only", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-05-21", "source": {"authors": ["Bangbang Zhou", "Hangdi Xing", "Yifan Chen", "Jianjun Xu", "Qi Zheng", "Feiyu Gao", "Zhibo Yang", "Shuai Bai", "Ming Yan", "Jieping Ye", "Hongtao Xie"], "categories": ["cs.AI"], "id": "2605.22100", "title": "MPDocBench-Parse: Benchmarking Practical Multi-page Document Parsing", "type": "arxiv", "url": "https://arxiv.org/abs/2605.22100"}, "sourceUpdatedAt": "2026-05-28T08:19:59Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "The benchmark addresses the lack of realistic multi-page document parsing evaluation, offering a more comprehensive protocol for assessing semantic continuity, hierarchical structure recovery, and visual content preservation, which is crucial for advancing document parsing systems in practical applications."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
