{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.19653", "kind": "paper", "url": "https://arxiv.org/abs/2608.19653"}, {"id": "gh:algorithmicresearchgroup/deltaml-bench-public", "kind": "repo", "url": "https://github.com/AlgorithmicResearchGroup/deltaml-bench-public"}], "authors": ["Josias Moukpe", "Priyanka Aryal", "Matthew Kenney"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.19653", "source_url": "https://arxiv.org/abs/2608.19653"}, "categories": ["General AI", "Language & Knowledge", "cs.LG"], "description": {"en": "DeltaML-Bench contains 48 tasks from research papers requiring agents to improve published baselines within open-source repositories, evaluated under compute constraints with fixed grading and integrity checks."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.19653", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.19653", "title": "DeltaML-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "key": "claire-radar:2608.19653", "modality": null, "name": "DeltaML-Bench", "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": "false", "origin_record_id": "bm_deltaml-bench_b6518e0f", "origin_source": "arxiv", "recognition_confidence": "1.0", "record_sha256": "dfd0eee03c4399b43ea841507321817bf726d9f18f26541d6077d21f0677ae96", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-name-audit-deferred", "reviewed_at": "2026-08-25T15:38:21.596820Z", "source_url": "https://arxiv.org/abs/2608.19653"}, "publisher": null, "released": "2026-08-20", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.19653", "source_url": "https://arxiv.org/abs/2608.19653"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-19653", "source": "claire_radar", "source_benchmark_id": "2608.19653", "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": "https://github.com/AlgorithmicResearchGroup/deltaml-bench-public", "githubScope": "benchmark_repo", "githubStars": 0, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2608.19653", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-10-04T15:31:48.030916Z", "state": "unavailable"}}, "status": "fresh"}, "capabilities": [], "confidence": "High", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-25T16:23:14.196185Z", "inputHash": "1cf16b69563516963cfef761bd4cd5c6b1a32eae53bed6a0f2bda2af3ddb8a12", "model": "deepseek-v4-flash"}, "curation": {"canonicalNameEvidence": "DeltaML-Bench: Evaluating Machine Learning Agents on Real-World Research Repositories", "canonicalNameSource": "paper_title", "decisionReason": "Unclear official benchmark name: paper uses DeltaML-Bench but repository is named DeltaMLBench; formal release needs confirmation from paper or README.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-25T15:38:21.596820Z", "state": "ai-name-audit-deferred"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "DeltaML-Bench contains 48 tasks from research papers requiring agents to improve published baselines within open-source repositories, evaluated under compute constraints with fixed grading and integrity checks.", "displayEligible": false, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["coined title prefix ending in Bench or Benchmark", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We introduce DeltaML-Bench, a benchmark comprising 48 tasks sourced from research papers that require agents to improve published baselines within imperfect, open-source repositories."}, "familyId": "bmf_dcd3591b7e50", "firstSeenAt": "2026-08-21", "heat": null, "id": "bm_deltaml-bench_b6518e0f", "indexedAt": "2026-08-21T04:30:09.569171Z", "industrySectors": [], "links": {"code": "https://github.com/AlgorithmicResearchGroup/deltaml-bench-public", "data": null, "paper": "https://arxiv.org/abs/2608.19653", "pdf": "https://arxiv.org/pdf/2608.19653", "project": null, "report": "https://arxiv.org/abs/2608.19653"}, "metrics": [], "motivation": "Autonomous agents for machine learning experimentation must navigate heterogeneous repositories, repair training pipelines, and evaluate candidate improvements under realistic compute constraints.", "name": "DeltaML-Bench", "oneLine": "DeltaML-Bench contains 48 tasks from research papers requiring agents to improve published baselines within open-source repositories, evaluated under compute constraints with fixed grading and integrity checks.", "paperTitle": "DeltaML-Bench: Evaluating Machine Learning Agents on Real-World Research Repositories", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.19653", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-21T04:30:09.569171Z"}, "publications": [], "publishers": [{"name": "Algorithmic Research Group", "organizationType": "benchmark-organization", "role": "benchmark-publisher", "sourceUrl": "https://github.com/AlgorithmicResearchGroup/deltaml-bench-public"}], "ranking": {"90d": {"confidence": "Low", "coverage": 0.55, "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": 0.0, "value": 0}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": null, "value": null}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": null, "value": null}}, "confidence": "Low", "coverage": 0.55, "rank": 521, "score": 23}, "method": "attention-ranking-v13", "rank": 521, "score": 23}}, "readiness": "Runnable", "recognitionConfidence": 1.0, "relation": "introduces", "releasedAt": "2026-08-20", "source": {"authors": ["Josias Moukpe", "Priyanka Aryal", "Matthew Kenney"], "categories": ["cs.LG", "cs.AI"], "id": "2608.19653", "title": "DeltaML-Bench: Evaluating Machine Learning Agents on Real-World Research Repositories", "type": "arxiv", "url": "https://arxiv.org/abs/2608.19653"}, "sourceUpdatedAt": "2026-08-20T05:42:39Z", "topics": ["cs.LG"], "venueAttempts": [], "whyItMatters": "It captures realistic conditions for autonomous ML experimentation, including heterogeneous codebases and specification gaming risks, providing a more faithful testbed for agent evaluation."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
