{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:4RBGFIJYXBEAZU2ATZZEMRY4OJ","short_pith_number":"pith:4RBGFIJY","schema_version":"1.0","canonical_sha256":"e44262a138b8480cd3409e7246471c724ad8655ff5fcfe20d70bcca92cae67f2","source":{"kind":"arxiv","id":"2108.08600","version":1},"attestation_state":"computed","paper":{"title":"Semantic Compositional Learning for Low-shot Scene Graph Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jianfei Cai, Jingkuan Song, Lianli Gao, Tao He, Yuan-Fang Li","submitted_at":"2021-08-19T10:13:55Z","abstract_excerpt":"Scene graphs provide valuable information to many downstream tasks. Many scene graph generation (SGG) models solely use the limited annotated relation triples for training, leading to their underperformance on low-shot (few and zero) scenarios, especially on the rare predicates. To address this problem, we propose a novel semantic compositional learning strategy that makes it possible to construct additional, realistic relation triples with objects from different images. Specifically, our strategy decomposes a relation triple by identifying and removing the unessential component and composes a"},"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":"2108.08600","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2021-08-19T10:13:55Z","cross_cats_sorted":[],"title_canon_sha256":"cccc1713406b5460d08887e153be2247e782d7dcdd45ca0f182b9600dbb8865b","abstract_canon_sha256":"d1e064ec05c1d088afbc38bf7ce1982bc1ab45f7953aab996b74f20941cdcb86"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:07:11.659928Z","signature_b64":"JfI2S3clZHvEHVimzB5oJgj+EtGM0yeTAKaSqMWfvD0sVFMXf0YBSPppz2BqCMsbMQhbDfzDE3haKagZAhmaAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e44262a138b8480cd3409e7246471c724ad8655ff5fcfe20d70bcca92cae67f2","last_reissued_at":"2026-07-05T03:07:11.659575Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:07:11.659575Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Semantic Compositional Learning for Low-shot Scene Graph Generation","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Jianfei Cai, Jingkuan Song, Lianli Gao, Tao He, Yuan-Fang Li","submitted_at":"2021-08-19T10:13:55Z","abstract_excerpt":"Scene graphs provide valuable information to many downstream tasks. Many scene graph generation (SGG) models solely use the limited annotated relation triples for training, leading to their underperformance on low-shot (few and zero) scenarios, especially on the rare predicates. To address this problem, we propose a novel semantic compositional learning strategy that makes it possible to construct additional, realistic relation triples with objects from different images. Specifically, our strategy decomposes a relation triple by identifying and removing the unessential component and composes a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2108.08600","kind":"arxiv","version":1},"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/2108.08600/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":"2108.08600","created_at":"2026-07-05T03:07:11.659631+00:00"},{"alias_kind":"arxiv_version","alias_value":"2108.08600v1","created_at":"2026-07-05T03:07:11.659631+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2108.08600","created_at":"2026-07-05T03:07:11.659631+00:00"},{"alias_kind":"pith_short_12","alias_value":"4RBGFIJYXBEA","created_at":"2026-07-05T03:07:11.659631+00:00"},{"alias_kind":"pith_short_16","alias_value":"4RBGFIJYXBEAZU2A","created_at":"2026-07-05T03:07:11.659631+00:00"},{"alias_kind":"pith_short_8","alias_value":"4RBGFIJY","created_at":"2026-07-05T03:07:11.659631+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.22274","citing_title":"CAGE-SGG: Counterfactual Active Graph Evidence for Open-Vocabulary Scene Graph Generation","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22546","citing_title":"ReLIC-SGG: Relation Lattice Completion for Open-Vocabulary Scene Graph Generation","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05057","citing_title":"ScriptHOI: Learning Scripted State Transitions for Open-Vocabulary Human-Object Interaction Detection","ref_index":58,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20823","citing_title":"RelWitness: Open-Vocabulary 3D Scene Graph Generation with Visual-Geometric Relation Witnesses","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2605.24634","citing_title":"Resolving Ambiguity in Composed Image Retrieval via Calibrated Interaction","ref_index":47,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20823","citing_title":"RelWitness: Open-Vocabulary 3D Scene Graph Generation with Visual-Geometric Relation Witnesses","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20606","citing_title":"Mind Your Margin and Boundary: Are Your Distilled Datasets Truly Robust?","ref_index":42,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20823","citing_title":"RelWitness: Open-Vocabulary 3D Scene Graph Generation with Visual-Geometric Relation Witnesses","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22546","citing_title":"ReLIC-SGG: Relation Lattice Completion for Open-Vocabulary Scene Graph Generation","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22274","citing_title":"CAGE-SGG: Counterfactual Active Graph Evidence for Open-Vocabulary Scene Graph Generation","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05057","citing_title":"ScriptHOI: Learning Scripted State Transitions for Open-Vocabulary Human-Object Interaction Detection","ref_index":58,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22546","citing_title":"ReLIC-SGG: Relation Lattice Completion for Open-Vocabulary Scene Graph Generation","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22274","citing_title":"CAGE-SGG: Counterfactual Active Graph Evidence for Open-Vocabulary Scene Graph Generation","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.05057","citing_title":"ScriptHOI: Learning Scripted State Transitions for Open-Vocabulary Human-Object Interaction Detection","ref_index":52,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4RBGFIJYXBEAZU2ATZZEMRY4OJ","json":"https://pith.science/pith/4RBGFIJYXBEAZU2ATZZEMRY4OJ.json","graph_json":"https://pith.science/api/pith-number/4RBGFIJYXBEAZU2ATZZEMRY4OJ/graph.json","events_json":"https://pith.science/api/pith-number/4RBGFIJYXBEAZU2ATZZEMRY4OJ/events.json","paper":"https://pith.science/paper/4RBGFIJY"},"agent_actions":{"view_html":"https://pith.science/pith/4RBGFIJYXBEAZU2ATZZEMRY4OJ","download_json":"https://pith.science/pith/4RBGFIJYXBEAZU2ATZZEMRY4OJ.json","view_paper":"https://pith.science/paper/4RBGFIJY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2108.08600&json=true","fetch_graph":"https://pith.science/api/pith-number/4RBGFIJYXBEAZU2ATZZEMRY4OJ/graph.json","fetch_events":"https://pith.science/api/pith-number/4RBGFIJYXBEAZU2ATZZEMRY4OJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4RBGFIJYXBEAZU2ATZZEMRY4OJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4RBGFIJYXBEAZU2ATZZEMRY4OJ/action/storage_attestation","attest_author":"https://pith.science/pith/4RBGFIJYXBEAZU2ATZZEMRY4OJ/action/author_attestation","sign_citation":"https://pith.science/pith/4RBGFIJYXBEAZU2ATZZEMRY4OJ/action/citation_signature","submit_replication":"https://pith.science/pith/4RBGFIJYXBEAZU2ATZZEMRY4OJ/action/replication_record"}},"created_at":"2026-07-05T03:07:11.659631+00:00","updated_at":"2026-07-05T03:07:11.659631+00:00"}