{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2NYDJU6CRS723RBLB2DUZX5IBH","short_pith_number":"pith:2NYDJU6C","schema_version":"1.0","canonical_sha256":"d37034d3c28cbfadc42b0e874cdfa809c5cb351f99d99bf54391cb1a9b3ea471","source":{"kind":"arxiv","id":"2607.00374","version":1},"attestation_state":"computed","paper":{"title":"Learning to Compose: Revisiting Proxy Task Design for Zero-Shot Composed Image Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.IR","cs.MM"],"primary_cat":"cs.CV","authors_text":"Jingjing Zhang, Lei Zhang, Zhendong Mao, Zheren Fu","submitted_at":"2026-07-01T03:20:06Z","abstract_excerpt":"Composed Image Retrieval (CIR) retrieves a target image from a reference image and a textual modification. While supervised CIR relies on costly triplets, Zero-Shot CIR (ZS-CIR) alleviates this reliance through proxy tasks trained on image-text pairs. However, existing proxy tasks primarily enhance visual and textual representations to accommodate a predefined composition mechanism such as pseudo-word injection into a frozen text encoder or linear feature arithmetic. As a result, the composition function itself remains unlearned, limiting the model's ability to express diverse and fine-grained"},"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":"2607.00374","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2026-07-01T03:20:06Z","cross_cats_sorted":["cs.AI","cs.CL","cs.IR","cs.MM"],"title_canon_sha256":"287ec6ba0f3b507af594a8939bd7c4a7c978df8e1ae377228ea81266a5b82c3c","abstract_canon_sha256":"1dc9de2c15cf3ef911426b3929f9870ba5303f3252d6ddeae10c34117087cbdc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-02T01:17:41.572527Z","signature_b64":"WkJN7qr/OY/kLIq/vbOUYPtQmm2qANT7Ym7Eg/VBnw8zVCNEkTvxifUVh1zfbxL/0StGGI/d/1NSesJv5wNNAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d37034d3c28cbfadc42b0e874cdfa809c5cb351f99d99bf54391cb1a9b3ea471","last_reissued_at":"2026-07-02T01:17:41.572074Z","signature_status":"signed_v1","first_computed_at":"2026-07-02T01:17:41.572074Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning to Compose: Revisiting Proxy Task Design for Zero-Shot Composed Image Retrieval","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.IR","cs.MM"],"primary_cat":"cs.CV","authors_text":"Jingjing Zhang, Lei Zhang, Zhendong Mao, Zheren Fu","submitted_at":"2026-07-01T03:20:06Z","abstract_excerpt":"Composed Image Retrieval (CIR) retrieves a target image from a reference image and a textual modification. While supervised CIR relies on costly triplets, Zero-Shot CIR (ZS-CIR) alleviates this reliance through proxy tasks trained on image-text pairs. However, existing proxy tasks primarily enhance visual and textual representations to accommodate a predefined composition mechanism such as pseudo-word injection into a frozen text encoder or linear feature arithmetic. As a result, the composition function itself remains unlearned, limiting the model's ability to express diverse and fine-grained"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.00374","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/2607.00374/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":"2607.00374","created_at":"2026-07-02T01:17:41.572131+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.00374v1","created_at":"2026-07-02T01:17:41.572131+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.00374","created_at":"2026-07-02T01:17:41.572131+00:00"},{"alias_kind":"pith_short_12","alias_value":"2NYDJU6CRS72","created_at":"2026-07-02T01:17:41.572131+00:00"},{"alias_kind":"pith_short_16","alias_value":"2NYDJU6CRS723RBL","created_at":"2026-07-02T01:17:41.572131+00:00"},{"alias_kind":"pith_short_8","alias_value":"2NYDJU6C","created_at":"2026-07-02T01:17:41.572131+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2NYDJU6CRS723RBLB2DUZX5IBH","json":"https://pith.science/pith/2NYDJU6CRS723RBLB2DUZX5IBH.json","graph_json":"https://pith.science/api/pith-number/2NYDJU6CRS723RBLB2DUZX5IBH/graph.json","events_json":"https://pith.science/api/pith-number/2NYDJU6CRS723RBLB2DUZX5IBH/events.json","paper":"https://pith.science/paper/2NYDJU6C"},"agent_actions":{"view_html":"https://pith.science/pith/2NYDJU6CRS723RBLB2DUZX5IBH","download_json":"https://pith.science/pith/2NYDJU6CRS723RBLB2DUZX5IBH.json","view_paper":"https://pith.science/paper/2NYDJU6C","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.00374&json=true","fetch_graph":"https://pith.science/api/pith-number/2NYDJU6CRS723RBLB2DUZX5IBH/graph.json","fetch_events":"https://pith.science/api/pith-number/2NYDJU6CRS723RBLB2DUZX5IBH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2NYDJU6CRS723RBLB2DUZX5IBH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2NYDJU6CRS723RBLB2DUZX5IBH/action/storage_attestation","attest_author":"https://pith.science/pith/2NYDJU6CRS723RBLB2DUZX5IBH/action/author_attestation","sign_citation":"https://pith.science/pith/2NYDJU6CRS723RBLB2DUZX5IBH/action/citation_signature","submit_replication":"https://pith.science/pith/2NYDJU6CRS723RBLB2DUZX5IBH/action/replication_record"}},"created_at":"2026-07-02T01:17:41.572131+00:00","updated_at":"2026-07-02T01:17:41.572131+00:00"}