{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2025:JPFAHTI7MC2A72NR6ASU6UAQ5I","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"2cca751c1741b89e7e6a8c11d162350b36619dfa93848a5e64d250175df44965","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-08-14T18:23:09Z","title_canon_sha256":"cd76d9e374b73fbc48c84652a0e6309c8acb83e4972e8c40d5a023fa9793350b"},"schema_version":"1.0","source":{"id":"2508.11011","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2508.11011","created_at":"2026-05-28T01:04:29Z"},{"alias_kind":"arxiv_version","alias_value":"2508.11011v2","created_at":"2026-05-28T01:04:29Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.11011","created_at":"2026-05-28T01:04:29Z"},{"alias_kind":"pith_short_12","alias_value":"JPFAHTI7MC2A","created_at":"2026-05-28T01:04:29Z"},{"alias_kind":"pith_short_16","alias_value":"JPFAHTI7MC2A72NR","created_at":"2026-05-28T01:04:29Z"},{"alias_kind":"pith_short_8","alias_value":"JPFAHTI7","created_at":"2026-05-28T01:04:29Z"}],"graph_snapshots":[{"event_id":"sha256:b3d35d88a8d61ec6f02085e8f3b1386cdd32112191d1f3d9bf23f3142f538333","target":"graph","created_at":"2026-05-28T01:04:29Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Our subsequent evaluation of current state-of-the-art large pre-trained VLMs shows notable generalization abilities in zero-shot and few-shot settings, while additional training is needed to make them applicable to actual construction sites."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"The three annotated tasks (captioning, safety-rule VQA, and element grounding) are sufficient to evaluate whether a VLM can function as an effective construction safety inspector (abstract, paragraph 3)."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"Presents the ConstructionSite 10k dataset and evaluates pre-trained VLMs on zero-shot and few-shot construction safety inspection tasks."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"Pre-trained vision-language models generalize to construction safety tasks in zero- and few-shot settings but still require additional training for real sites."}],"snapshot_sha256":"daa266f771820fb89720aa256f46e2a8e8ed308d63d417a02fda9645db419e08"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"2c630594ba58b072b8518a04e48ce036930837e7de969f89284e2af4769e0ac8"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2508.11011/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Construction safety inspections typically involve a human inspector identifying safety concerns on-site. With the rise of powerful Vision Language Models (VLMs), researchers are exploring their use for tasks such as detecting safety rule violations from on-site images. However, there is a lack of open datasets to comprehensively evaluate and further fine-tune VLMs in construction safety inspection. Current applications of VLMs use small, supervised datasets, limiting their applicability in tasks they are not directly trained for. In this paper, we propose the ConstructionSite 10k, featuring 10","authors_text":"Xuezheng Chen, Zhengbo Zou","cross_cats":[],"headline":"Pre-trained vision-language models generalize to construction safety tasks in zero- and few-shot settings but still require additional training for real sites.","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-08-14T18:23:09Z","title":"Are Large Pre-trained Vision Language Models Effective Construction Safety Inspectors?"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.11011","kind":"arxiv","version":2},"verdict":{"created_at":"2026-05-18T22:27:29.248449Z","id":"ac577dbc-fa9a-4ee2-b832-dc50d7a3360e","model_set":{"reader":"grok-4.3"},"one_line_summary":"Presents the ConstructionSite 10k dataset and evaluates pre-trained VLMs on zero-shot and few-shot construction safety inspection tasks.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"Pre-trained vision-language models generalize to construction safety tasks in zero- and few-shot settings but still require additional training for real sites.","strongest_claim":"Our subsequent evaluation of current state-of-the-art large pre-trained VLMs shows notable generalization abilities in zero-shot and few-shot settings, while additional training is needed to make them applicable to actual construction sites.","weakest_assumption":"The three annotated tasks (captioning, safety-rule VQA, and element grounding) are sufficient to evaluate whether a VLM can function as an effective construction safety inspector (abstract, paragraph 3)."}},"verdict_id":"ac577dbc-fa9a-4ee2-b832-dc50d7a3360e"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:393056edf4e6dfb7fd807171f351136047b3d05c808e78907e18e966e7daa994","target":"record","created_at":"2026-05-28T01:04:29Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"2cca751c1741b89e7e6a8c11d162350b36619dfa93848a5e64d250175df44965","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CV","submitted_at":"2025-08-14T18:23:09Z","title_canon_sha256":"cd76d9e374b73fbc48c84652a0e6309c8acb83e4972e8c40d5a023fa9793350b"},"schema_version":"1.0","source":{"id":"2508.11011","kind":"arxiv","version":2}},"canonical_sha256":"4bca03cd1f60b40fe9b1f0254f5010ea0cce5202b38a758ebfa2cf97e6ddd121","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"4bca03cd1f60b40fe9b1f0254f5010ea0cce5202b38a758ebfa2cf97e6ddd121","first_computed_at":"2026-05-28T01:04:29.625881Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-28T01:04:29.625881Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"2/AS4YNxgUhMKGGplLGGAOw8+MMmnfn95Fxi+plbONIs6lgsm8WGznA1UUMn8b7ZDEBHvWI0H2hurDeHrY9oBg==","signature_status":"signed_v1","signed_at":"2026-05-28T01:04:29.626436Z","signed_message":"canonical_sha256_bytes"},"source_id":"2508.11011","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:393056edf4e6dfb7fd807171f351136047b3d05c808e78907e18e966e7daa994","sha256:b3d35d88a8d61ec6f02085e8f3b1386cdd32112191d1f3d9bf23f3142f538333"],"state_sha256":"e6cc1c8887096e2d6d5d41f7696c097260a18f182661883190d24f5472bb35ed"}