{"record": {"aliases": [], "artifacts": [{"id": "arxiv:2608.10317", "kind": "paper", "url": "https://arxiv.org/abs/2608.10317"}, {"id": "hf:nvidia/physicalai-traffic-anomaly-reasoning", "kind": "dataset", "url": "https://huggingface.co/datasets/nvidia/PhysicalAI-Traffic-Anomaly-Reasoning"}], "authors": ["Han Zhang", "Yilin Zhao", "Zaid Pervaiz Bhat", "Zheng Tang", "Varun Praveen", "Vidya N. Murali", "David C. Anastasiu", "Tomasz Kornuta"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2608.10317", "source_url": "https://arxiv.org/abs/2608.10317"}, "categories": ["General AI", "Multimodal", "Reasoning"], "description": {"en": "Evaluates video-language models on ten traffic anomaly reasoning tasks using 960 human-curated test annotations over 80 held-out clips."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2608.10317", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2608.10317", "title": "TAR-Bench"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Santa Clara University"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2608.10317", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2608.10317", "source_url": "https://openalex.org/W7202256505"}, "key": "claire-radar:2608.10317", "modality": null, "name": "TAR-Bench", "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_tar-bench_36cf86db", "origin_source": "arxiv", "recognition_confidence": "0.7", "record_sha256": "845c03677f15933ce4a67f6c9f7aa523a256db56d31daff7a118a0c7f7e9ec49", "relation": "introduces", "review_model": "deepseek-v4-pro", "review_state": "ai-reviewed", "reviewed_at": "2026-08-27T04:14:52.836173Z", "source_url": "https://arxiv.org/abs/2608.10317"}, "publisher": null, "released": "2026-08-10", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2608.10317", "source_url": "https://arxiv.org/abs/2608.10317"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2608-10317", "source": "claire_radar", "source_benchmark_id": "2608.10317", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": [], "annotation": "Mixed", "applicationDomains": ["General AI"], "area": "Multimodal", "attention": {"asOf": "2026-10-05", "attemptedAt": "2026-10-04T15:31:48.030916Z", "githubRepo": null, "githubScope": null, "githubStars": null, "hfDailySubmittedAt": null, "hfDataset": "nvidia/PhysicalAI-Traffic-Anomaly-Reasoning", "hfDatasetDownloads": 1083, "hfDatasetLikes": 20, "hfPaperUpvotes": 0, "hfPaperUrl": "https://huggingface.co/papers/2608.10317", "observedAt": "2026-10-04T15:31:48.030916Z", "signalStatus": {"githubStars": {"state": "not_applicable"}, "hfDatasetDownloads": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfDatasetLikes": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}, "hfPaperUpvotes": {"observedAt": "2026-10-04T15:31:48.030916Z", "state": "fresh"}}, "status": "fresh"}, "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "The benchmark is tied to AI City Challenge 2026, has a public Hugging Face dataset, and addresses a practical traffic reasoning use case with broad model evaluation results.", "score": 70}, "capabilities": ["Reasoning"], "confidence": "Medium", "construction": "Aggregate Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-08-27T04:14:52.836173Z", "inputHash": "5d3718b514b69f7c93ac1d2b4eaa976bd9ca928047209fce9205b19de74448b3", "model": "deepseek-v4-pro"}, "curation": {"canonicalNameEvidence": "Its evaluation component, TAR-Bench, contains 960 human-curated test annotations for 80 held-out clips", "canonicalNameSource": "abstract", "decisionReason": "The benchmark has a defined evaluation set, scoring via aggregate score, public dataset with download script, and is used in an official challenge.", "model": "deepseek-v4-pro", "reviewedAt": "2026-08-27T04:14:52.836173Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "Evaluates video-language models on ten traffic anomaly reasoning tasks using 960 human-curated test annotations over 80 held-out clips.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "score_submission", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence", "public artifact URL"], "snippet": "We present TAR (Traffic Anomaly Reasoning) and TAR-Bench datasets, resources for training and evaluating video-language models beyond anomaly detection."}, "familyId": "bmf_484abfae69f2", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_tar-bench_36cf86db", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": "https://huggingface.co/datasets/nvidia/PhysicalAI-Traffic-Anomaly-Reasoning", "paper": "https://arxiv.org/abs/2608.10317", "pdf": "https://arxiv.org/pdf/2608.10317", "project": null, "report": "https://arxiv.org/abs/2608.10317"}, "metrics": [], "motivation": "We present TAR (Traffic Anomaly Reasoning) and TAR-Bench datasets, resources for training and evaluating video-language models beyond anomaly detection.", "name": "TAR-Bench", "oneLine": "Evaluates video-language models on ten traffic anomaly reasoning tasks using 960 human-curated test annotations over 80 held-out clips.", "paperTitle": "From Detection to Understanding: TAR and TAR-Bench for Multi-Task Traffic Anomaly Reasoning", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2608.10317", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "publishers": [{"name": "NVIDIA", "organizationType": "company-research-lab", "role": "benchmark-publisher", "sourceUrl": "https://huggingface.co/datasets/nvidia/PhysicalAI-Traffic-Anomaly-Reasoning"}], "ranking": {"90d": {"confidence": "Medium", "coverage": 0.45, "datasetDownloadRank": 17, "datasetRankPopulation": 85, "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": null, "value": null}, "hfDatasetDownloads": {"normalization": "log-max", "normalized": 0.7196662505600427, "value": 1083}, "hfPaperUpvotes": {"normalization": "log-max", "normalized": 0.0, "value": 0}}, "confidence": "Medium", "coverage": 0.45, "rank": 71, "score": 49}, "method": "attention-ranking-v13", "rank": 71, "score": 49}}, "readiness": "Inspectable", "recognitionConfidence": 0.7, "relation": "introduces", "releasedAt": "2026-08-10", "source": {"authors": ["Han Zhang", "Yilin Zhao", "Zaid Pervaiz Bhat", "Zheng Tang", "Varun Praveen", "Vidya N. Murali", "David C. Anastasiu", "Tomasz Kornuta"], "categories": ["cs.CV"], "id": "2608.10317", "title": "From Detection to Understanding: TAR and TAR-Bench for Multi-Task Traffic Anomaly Reasoning", "type": "arxiv", "url": "https://arxiv.org/abs/2608.10317"}, "sourceUpdatedAt": "2026-08-10T23:40:07Z", "topics": ["Reasoning"], "venueAttempts": [], "whyItMatters": "Fills the gap between anomaly detection and higher-level reasoning by testing temporal localization, causal understanding, and multi-task performance."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
