{"record": {"aliases": ["Field Order Should Not Matter: Permutation-Invariant Embedding Model Fine-Tuning for Structured Metadata Retrieval"], "artifacts": [{"id": "arxiv:2606.30473", "kind": "paper", "url": "https://arxiv.org/abs/2606.30473"}], "authors": ["Aivin V. Solatorio", "Olivier Dupriez", "Rafael Macalaba"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2606.30473", "source_url": "https://arxiv.org/abs/2606.30473"}, "categories": ["General AI", "Information retrieval", "Language & Knowledge", "cs.CL"], "description": {"en": "DevDataBench evaluates structured-metadata retrieval using grounded, facet-targeted queries across 15 languages."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2606.30473", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2606.30473", "title": "DevDataBench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["World Bank Group"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2606.30473", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2606.30473", "source_url": "https://openalex.org/W7166742385"}, "key": "claire-radar:2606.30473", "modality": null, "name": "DevDataBench", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_field-order-should-not-matter-permutation-_190c82ac", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "9cfb8c6a8dc897ee5e66e205d6c92322e9458854227444ca84688887c0bdc336", "relation": "introduces", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T18:55:07.505928Z", "source_url": "https://arxiv.org/abs/2606.30473"}, "publisher": null, "released": "2026-06-29", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2606.30473", "source_url": "https://arxiv.org/abs/2606.30473"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2606-30473", "source": "claire_radar", "source_benchmark_id": "2606.30473", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["Field Order Should Not Matter: Permutation-Invariant Embedding Model Fine-Tuning for Structured Metadata Retrieval"], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attentionForecast": {"confidence": "Low", "horizon": "7d", "reason": "The source supplies no attention signals; the forecast reflects only the benchmark's multilingual scope and stated public release.", "score": 3}, "capabilities": ["Information retrieval"], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T18:55:07.505928Z", "inputHash": "47e29798cbf5b644d23c11a43db42e47d8a09d2f5d929f404f056e13be8e0e58", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "$\\textbf{DevDataBench}$ is a fully LLM-generated benchmark", "canonicalNameSource": "abstract", "decisionReason": "The abstract formally names the benchmark, reports comparable nDCG@10 scoring, and states that the benchmark and supporting pipeline are released.", "evaluationEvidence": {"reuse": "We release the benchmark, pipeline, models, and a reusable PI-FT framework.", "scoring": "0.707 vs.\\ 0.556 nDCG@10", "task": "grounded, facet-targeted queries across 15 languages"}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T18:55:07.505928Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "DevDataBench evaluates structured-metadata retrieval using grounded, facet-targeted queries across 15 languages.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We release the benchmark, pipeline, models, and a reusable PI-FT framework."}, "familyId": "bmf_75fdeacfcf81", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_field-order-should-not-matter-permutation-_190c82ac", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2606.30473", "pdf": "https://arxiv.org/pdf/2606.30473", "project": null, "report": "https://arxiv.org/abs/2606.30473"}, "metrics": [], "motivation": "We study retrieval over catalogs of structured metadata, where each record is a small schema whose fields answer different kinds of query.", "name": "DevDataBench", "oneLine": "DevDataBench evaluates structured-metadata retrieval using grounded, facet-targeted queries across 15 languages.", "paperTitle": "Field Order Should Not Matter: Permutation-Invariant Embedding Model Fine-Tuning for Structured Metadata Retrieval", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2606.30473", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "readiness": "Paper only", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-06-29", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Aivin V. Solatorio", "Olivier Dupriez", "and Rafael Macalaba"], "categories": ["cs.CL", "cs.AI", "cs.IR", "cs.LG", "econ.GN", "q-fin.EC"], "id": "2606.30473", "title": "Field Order Should Not Matter: Permutation-Invariant Embedding Model Fine-Tuning for Structured Metadata Retrieval", "type": "arxiv", "url": "https://arxiv.org/abs/2606.30473"}, "sourceUpdatedAt": "2026-06-29T15:33:04Z", "topics": ["cs.CL"], "venueAttempts": [], "whyItMatters": "It measures multilingual discovery of development-statistics indicators while exposing retrieval sensitivity to field ordering."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
