{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:UN6BXNWPWW7BOFZGD6BGYQK2DR","short_pith_number":"pith:UN6BXNWP","schema_version":"1.0","canonical_sha256":"a37c1bb6cfb5be1717261f826c415a1c7a45bf4e9d7855e3ec1faba0511ee99e","source":{"kind":"arxiv","id":"2406.19934","version":2},"attestation_state":"computed","paper":{"title":"From the Least to the Most: Building a Plug-and-Play Visual Reasoner via Data Synthesis","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chuanqi Cheng, Jian Guan, Rui Yan, Wei Wu","submitted_at":"2024-06-28T14:04:10Z","abstract_excerpt":"We explore multi-step reasoning in vision-language models (VLMs). The problem is challenging, as reasoning data consisting of multiple steps of visual and language processing are barely available. To overcome the challenge, we first introduce a least-to-most visual reasoning paradigm, which interleaves steps of decomposing a question into sub-questions and invoking external tools for resolving sub-questions. Based on the paradigm, we further propose a novel data synthesis approach that can automatically create questions and multi-step reasoning paths for an image in a bottom-up manner. Our app"},"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":"2406.19934","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.CL","submitted_at":"2024-06-28T14:04:10Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"23a0a2a9aa2b46833c317c43de0f669ed242bb5cec6319a2290962c0dd725c29","abstract_canon_sha256":"1f27c39d8c451fc7f2b8846dd4e2e5ed0a6e1ad9f8eddf2a8e32fc3547a9f1af"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:19:10.962468Z","signature_b64":"F8Grry5q/O8RRKuP3VlBrA9+1hTtN0Wc064WV9QNXDl5omPPUA9MLQkuwU+Zv0y3BFTYBsIMCrdaX8HqarTQCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a37c1bb6cfb5be1717261f826c415a1c7a45bf4e9d7855e3ec1faba0511ee99e","last_reissued_at":"2026-07-05T09:19:10.962009Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:19:10.962009Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"From the Least to the Most: Building a Plug-and-Play Visual Reasoner via Data Synthesis","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Chuanqi Cheng, Jian Guan, Rui Yan, Wei Wu","submitted_at":"2024-06-28T14:04:10Z","abstract_excerpt":"We explore multi-step reasoning in vision-language models (VLMs). The problem is challenging, as reasoning data consisting of multiple steps of visual and language processing are barely available. To overcome the challenge, we first introduce a least-to-most visual reasoning paradigm, which interleaves steps of decomposing a question into sub-questions and invoking external tools for resolving sub-questions. Based on the paradigm, we further propose a novel data synthesis approach that can automatically create questions and multi-step reasoning paths for an image in a bottom-up manner. Our app"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.19934","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/2406.19934/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":"2406.19934","created_at":"2026-07-05T09:19:10.962065+00:00"},{"alias_kind":"arxiv_version","alias_value":"2406.19934v2","created_at":"2026-07-05T09:19:10.962065+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.19934","created_at":"2026-07-05T09:19:10.962065+00:00"},{"alias_kind":"pith_short_12","alias_value":"UN6BXNWPWW7B","created_at":"2026-07-05T09:19:10.962065+00:00"},{"alias_kind":"pith_short_16","alias_value":"UN6BXNWPWW7BOFZG","created_at":"2026-07-05T09:19:10.962065+00:00"},{"alias_kind":"pith_short_8","alias_value":"UN6BXNWP","created_at":"2026-07-05T09:19:10.962065+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2504.20199","citing_title":"Weaving Context Across Images: Improving Vision-Language Models through Focus-Centric Visual Chains","ref_index":9,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UN6BXNWPWW7BOFZGD6BGYQK2DR","json":"https://pith.science/pith/UN6BXNWPWW7BOFZGD6BGYQK2DR.json","graph_json":"https://pith.science/api/pith-number/UN6BXNWPWW7BOFZGD6BGYQK2DR/graph.json","events_json":"https://pith.science/api/pith-number/UN6BXNWPWW7BOFZGD6BGYQK2DR/events.json","paper":"https://pith.science/paper/UN6BXNWP"},"agent_actions":{"view_html":"https://pith.science/pith/UN6BXNWPWW7BOFZGD6BGYQK2DR","download_json":"https://pith.science/pith/UN6BXNWPWW7BOFZGD6BGYQK2DR.json","view_paper":"https://pith.science/paper/UN6BXNWP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2406.19934&json=true","fetch_graph":"https://pith.science/api/pith-number/UN6BXNWPWW7BOFZGD6BGYQK2DR/graph.json","fetch_events":"https://pith.science/api/pith-number/UN6BXNWPWW7BOFZGD6BGYQK2DR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UN6BXNWPWW7BOFZGD6BGYQK2DR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UN6BXNWPWW7BOFZGD6BGYQK2DR/action/storage_attestation","attest_author":"https://pith.science/pith/UN6BXNWPWW7BOFZGD6BGYQK2DR/action/author_attestation","sign_citation":"https://pith.science/pith/UN6BXNWPWW7BOFZGD6BGYQK2DR/action/citation_signature","submit_replication":"https://pith.science/pith/UN6BXNWPWW7BOFZGD6BGYQK2DR/action/replication_record"}},"created_at":"2026-07-05T09:19:10.962065+00:00","updated_at":"2026-07-05T09:19:10.962065+00:00"}