{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:P4NTZRGMHK3V4QAMQTWNN3UDRR","short_pith_number":"pith:P4NTZRGM","schema_version":"1.0","canonical_sha256":"7f1b3cc4cc3ab75e400c84ecd6ee838c53a2188bd8594b4683874404bc0e5bac","source":{"kind":"arxiv","id":"2508.07996","version":2},"attestation_state":"computed","paper":{"title":"Structured Relational Reasoning for Group Activity Assessment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Alina Roitberg, Chengzheng Yang, Thinesh Thiyakesan Ponbagavathi","submitted_at":"2025-08-11T13:59:22Z","abstract_excerpt":"Group Activity Detection (GAD) involves recognizing social groups and their collective behaviors in videos. Vision Foundation Models (VFMs), like DINOv2, offer excellent features but are pretrained on object-centric data. We find that naively substituting them into existing GAD pipelines actually degrades performance, exposing structured group-aware decoding as the true bottleneck.\n  We introduce ProGraD, a structured relational-reasoning framework for GAD built on top of frozen VFMs. At its core is a lightweight two-layer GroupContext Transformer that explicitly models actor-group association"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2508.07996","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-08-11T13:59:22Z","cross_cats_sorted":[],"title_canon_sha256":"f47c973f1e41589455dfbe5dc2b8e0043df98b1ea2226ab70f6413925976dd38","abstract_canon_sha256":"6d5fc9aea8e10c8ce897dc115bb4d9707503a8662f7468189bb1c42ac066eaa1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-27T01:05:37.322492Z","signature_b64":"4B8td1gm79848a3LC/soJKj8MqOootXdb/KpOJg5dMG5UqJ1XC0KVWqxrf2+rZCcRRKu2C+bd8KDdm53G2n3AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7f1b3cc4cc3ab75e400c84ecd6ee838c53a2188bd8594b4683874404bc0e5bac","last_reissued_at":"2026-05-27T01:05:37.321754Z","signature_status":"signed_v1","first_computed_at":"2026-05-27T01:05:37.321754Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Structured Relational Reasoning for Group Activity Assessment","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Alina Roitberg, Chengzheng Yang, Thinesh Thiyakesan Ponbagavathi","submitted_at":"2025-08-11T13:59:22Z","abstract_excerpt":"Group Activity Detection (GAD) involves recognizing social groups and their collective behaviors in videos. Vision Foundation Models (VFMs), like DINOv2, offer excellent features but are pretrained on object-centric data. We find that naively substituting them into existing GAD pipelines actually degrades performance, exposing structured group-aware decoding as the true bottleneck.\n  We introduce ProGraD, a structured relational-reasoning framework for GAD built on top of frozen VFMs. At its core is a lightweight two-layer GroupContext Transformer that explicitly models actor-group association"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.07996","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2508.07996/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2508.07996","created_at":"2026-05-27T01:05:37.321850+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.07996v2","created_at":"2026-05-27T01:05:37.321850+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.07996","created_at":"2026-05-27T01:05:37.321850+00:00"},{"alias_kind":"pith_short_12","alias_value":"P4NTZRGMHK3V","created_at":"2026-05-27T01:05:37.321850+00:00"},{"alias_kind":"pith_short_16","alias_value":"P4NTZRGMHK3V4QAM","created_at":"2026-05-27T01:05:37.321850+00:00"},{"alias_kind":"pith_short_8","alias_value":"P4NTZRGM","created_at":"2026-05-27T01:05:37.321850+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.18421","citing_title":"Why Relational Graphs Will Save the Next Generation of Vision Foundation Models?","ref_index":19,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/P4NTZRGMHK3V4QAMQTWNN3UDRR","json":"https://pith.science/pith/P4NTZRGMHK3V4QAMQTWNN3UDRR.json","graph_json":"https://pith.science/api/pith-number/P4NTZRGMHK3V4QAMQTWNN3UDRR/graph.json","events_json":"https://pith.science/api/pith-number/P4NTZRGMHK3V4QAMQTWNN3UDRR/events.json","paper":"https://pith.science/paper/P4NTZRGM"},"agent_actions":{"view_html":"https://pith.science/pith/P4NTZRGMHK3V4QAMQTWNN3UDRR","download_json":"https://pith.science/pith/P4NTZRGMHK3V4QAMQTWNN3UDRR.json","view_paper":"https://pith.science/paper/P4NTZRGM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.07996&json=true","fetch_graph":"https://pith.science/api/pith-number/P4NTZRGMHK3V4QAMQTWNN3UDRR/graph.json","fetch_events":"https://pith.science/api/pith-number/P4NTZRGMHK3V4QAMQTWNN3UDRR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/P4NTZRGMHK3V4QAMQTWNN3UDRR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/P4NTZRGMHK3V4QAMQTWNN3UDRR/action/storage_attestation","attest_author":"https://pith.science/pith/P4NTZRGMHK3V4QAMQTWNN3UDRR/action/author_attestation","sign_citation":"https://pith.science/pith/P4NTZRGMHK3V4QAMQTWNN3UDRR/action/citation_signature","submit_replication":"https://pith.science/pith/P4NTZRGMHK3V4QAMQTWNN3UDRR/action/replication_record"}},"created_at":"2026-05-27T01:05:37.321850+00:00","updated_at":"2026-05-27T01:05:37.321850+00:00"}