{"record": {"aliases": ["ExPhy: A Benchmark for Explicit Physical Property Learning in Multi-Object Trajectory Forecasting"], "artifacts": [{"id": "arxiv:2608.20009", "kind": "paper", "url": "https://arxiv.org/abs/2608.20009"}, {"id": "gh:zest86/exphy", "kind": "repo", "url": "https://github.com/Zest86/ExPhy"}], "authors": ["Rui Wang", "Yeteng Wu", "Xianlin Zhang", "Mengshi Qi"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.20009", "source_url": "https://arxiv.org/abs/2608.20009"}, "categories": ["General AI", "Language & Knowledge", "cs.AI"], "description": {"en": "24,000 simulated scenes for multi-object trajectory forecasting with explicit labels for mass, friction, and restitution, including ID and OOD splits over physical parameters and initial states."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.20009", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.20009", "title": "ExPhy"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Beijing University of Posts and Telecommunications", "State Key Laboratory of Networking and Switching Technology"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.20009", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.20009", "source_url": "https://openalex.org/W7203874793"}, "key": "claire-radar:2608.20009", "modality": null, "name": "ExPhy", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-10-14T09:44:51Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_exphy-a-benchmark-for-explicit-physical-pr_ac5ba09f", "origin_source": "arxiv", "recognition_confidence": "0.7", "record_sha256": "9f58e095efff97ba5b2f734af22a1781be5d136a23bb3a85b6cb2a9c9c67b092", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-08-26T06:15:55.708965Z", "source_url": "https://arxiv.org/abs/2608.20009"}, "publisher": null, "released": "2026-08-20", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.20009", "source_url": "https://arxiv.org/abs/2608.20009"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-20009", "source": "claire_radar", "source_benchmark_id": "2608.20009", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["ExPhy: A Benchmark for Explicit Physical Property Learning in Multi-Object Trajectory Forecasting"], "annotation": "Unknown", "applicationDomains": ["General AI"], "area": "Language & Knowledge", "attention": {"asOf": "2026-08-26", "attemptedAt": "2026-08-25T23:36:26.005677Z", "githubRepo": "https://github.com/Zest86/ExPhy", "githubScope": "benchmark_repo", "githubStars": 0, "hfDailySubmittedAt": null, "hfDataset": null, "hfDatasetDownloads": null, "hfDatasetLikes": null, "hfPaperUpvotes": null, "hfPaperUrl": "https://huggingface.co/papers/2608.20009", "observedAt": "2026-08-25T23:36:26.005677Z", "signalStatus": {"githubStars": {"observedAt": "2026-08-25T23:36:26.005677Z", "state": "fresh"}, "hfDatasetDownloads": {"state": "not_applicable"}, "hfDatasetLikes": {"state": "not_applicable"}, "hfPaperUpvotes": {"checkedAt": "2026-08-25T23:36:26.005677Z", "state": "unavailable"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "The benchmark addresses an emerging area of physics-guided learning and provides open code and data, likely to receive focused attention from trajectory forecasting researchers.", "score": 40}, "capabilities": [], "confidence": "Medium", "construction": "Unknown", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-26T06:15:55.708965Z", "inputHash": "a25cfacd150c6c3ba74a85bae4812c3cdde3401eaa8774be6430686559d7f76c", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "ExPhy: A Benchmark for Explicit Physical Property Learning in Multi-Object Trajectory Forecasting", "canonicalNameSource": "paper_title", "decisionReason": "The paper clearly defines the benchmark and provides a GitHub repository with code and data, ensuring reproducibility and public reuse.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-26T06:15:55.708965Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "24,000 simulated scenes for multi-object trajectory forecasting with explicit labels for mass, friction, and restitution, including ID and OOD splits over physical parameters and initial states.", "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", "public artifact URL"], "snippet": "To address this gap, we introduce \\emph{ExPhy}, a multi-object trajectory forecasting benchmark containing 24,000 simulated physical scenes with explicit object-level labels for mass, friction, and restitution."}, "familyId": "bmf_1779aacd1819", "firstSeenAt": "2026-08-21", "heat": null, "id": "bm_exphy-a-benchmark-for-explicit-physical-pr_ac5ba09f", "indexedAt": "2026-08-21T04:30:09.569171Z", "industrySectors": [], "links": {"code": "https://github.com/Zest86/ExPhy", "data": null, "paper": "https://arxiv.org/abs/2608.20009", "pdf": "https://arxiv.org/pdf/2608.20009", "project": null, "report": "https://arxiv.org/abs/2608.20009"}, "metrics": [], "motivation": "Understanding object dynamics requires not only predicting future trajectories but also examining whether a model captures the physical properties that govern motion.", "name": "ExPhy", "oneLine": "24,000 simulated scenes for multi-object trajectory forecasting with explicit labels for mass, friction, and restitution, including ID and OOD splits over physical parameters and initial states.", "paperTitle": "ExPhy: A Benchmark for Explicit Physical Property Learning in Multi-Object Trajectory Forecasting", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.20009", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-21T04:30:09.569171Z"}, "publications": [], "publishers": [{"name": "Beijing University of Posts and Telecommunications", "organizationType": "academic-lab", "role": "benchmark-publisher", "sourceUrl": "https://github.com/Zest86/ExPhy"}], "ranking": {"30d": {"confidence": "Low", "coverage": 0.55, "growth": {"components": {"githubStars": {"percentile": null, "value": null}, "hfDatasetDownloads": {"percentile": null, "value": null}, "hfPaperUpvotes": {"percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 30}, "level": {"components": {"githubStars": {"percentile": 0.18928571428571428, "value": 0}, "hfDatasetDownloads": {"percentile": null, "value": null}, "hfPaperUpvotes": {"percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.55, "rank": 133, "score": 33}, "method": "attention-ranking-v11", "rank": 133, "score": 33}, "90d": {"confidence": "Low", "coverage": 0.55, "growth": {"components": {"githubStars": {"percentile": null, "value": null}, "hfDatasetDownloads": {"percentile": null, "value": null}, "hfPaperUpvotes": {"percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.0, "elapsedDays": null, "historyDate": null, "rank": null, "score": null, "windowDays": 90}, "level": {"components": {"githubStars": {"percentile": 0.11716621253405994, "value": 0}, "hfDatasetDownloads": {"percentile": null, "value": null}, "hfPaperUpvotes": {"percentile": null, "value": null}}, "confidence": "Low", "coverage": 0.55, "rank": 346, "score": 29}, "method": "attention-ranking-v11", "rank": 346, "score": 29}}, "readiness": "Runnable", "recognitionConfidence": 0.7, "relation": "introduces", "releasedAt": "2026-08-20", "source": {"authors": ["Rui Wang", "Yeteng Wu", "Xianlin Zhang", "Mengshi Qi"], "categories": ["cs.AI"], "id": "2608.20009", "title": "ExPhy: A Benchmark for Explicit Physical Property Learning in Multi-Object Trajectory Forecasting", "type": "arxiv", "url": "https://arxiv.org/abs/2608.20009"}, "sourceUpdatedAt": "2026-08-20T13:24:13Z", "topics": ["cs.AI"], "venueAttempts": [], "whyItMatters": "Enables joint evaluation of trajectory forecasting and physical property estimation, revealing that accurate forecasts do not guarantee correct physical understanding."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
