{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:CAH45CJZEXVLNYYSISZLVA53PL","short_pith_number":"pith:CAH45CJZ","schema_version":"1.0","canonical_sha256":"100fce893925eab6e31244b2ba83bb7adabadde223642574453cc8039c4893f8","source":{"kind":"arxiv","id":"2505.17609","version":2},"attestation_state":"computed","paper":{"title":"Integrating Visual Interpretation and Linguistic Reasoning for Math Problem Solving","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jinfeng Bai, Lei Zhang, Ming Liu, Qilong Wang, Wangmeng Zuo, Zhilong Ji, Zixian Guo","submitted_at":"2025-05-23T08:18:00Z","abstract_excerpt":"Current large vision-language models (LVLMs) typically employ a connector module to link visual features with text embeddings of large language models (LLMs) and use end-to-end training to achieve multi-modal understanding in a unified process. Effective alignment needs high-quality pre-training data and a carefully designed training process. Current LVLMs face challenges when addressing complex vision-language reasoning tasks, with their reasoning capabilities notably lagging behind those of LLMs. This paper proposes a paradigm shift: instead of training end-to-end vision-language reasoning m"},"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":"2505.17609","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-05-23T08:18:00Z","cross_cats_sorted":[],"title_canon_sha256":"cfa1c72cdf61b12aa172132dbfb35f966ad9323bd89105994fa69e6b141e3fd8","abstract_canon_sha256":"91938c141c6550be5dd82d46b47c27aa08563ecbf34ec617820ce2eefb8f04f2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:53:03.368816Z","signature_b64":"+kAOfC+mUNqSq/63niZtEiq/4jMfNeiPgAJt0CDf8Mvm9cTv9oohGc06Cgrwu2HGSWjOqpkI8UBvZhHO4uayAw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"100fce893925eab6e31244b2ba83bb7adabadde223642574453cc8039c4893f8","last_reissued_at":"2026-07-05T11:53:03.368334Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:53:03.368334Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Integrating Visual Interpretation and Linguistic Reasoning for Math Problem Solving","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Jinfeng Bai, Lei Zhang, Ming Liu, Qilong Wang, Wangmeng Zuo, Zhilong Ji, Zixian Guo","submitted_at":"2025-05-23T08:18:00Z","abstract_excerpt":"Current large vision-language models (LVLMs) typically employ a connector module to link visual features with text embeddings of large language models (LLMs) and use end-to-end training to achieve multi-modal understanding in a unified process. Effective alignment needs high-quality pre-training data and a carefully designed training process. Current LVLMs face challenges when addressing complex vision-language reasoning tasks, with their reasoning capabilities notably lagging behind those of LLMs. This paper proposes a paradigm shift: instead of training end-to-end vision-language reasoning m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.17609","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/2505.17609/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":"2505.17609","created_at":"2026-07-05T11:53:03.368392+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.17609v2","created_at":"2026-07-05T11:53:03.368392+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.17609","created_at":"2026-07-05T11:53:03.368392+00:00"},{"alias_kind":"pith_short_12","alias_value":"CAH45CJZEXVL","created_at":"2026-07-05T11:53:03.368392+00:00"},{"alias_kind":"pith_short_16","alias_value":"CAH45CJZEXVLNYYS","created_at":"2026-07-05T11:53:03.368392+00:00"},{"alias_kind":"pith_short_8","alias_value":"CAH45CJZ","created_at":"2026-07-05T11:53:03.368392+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2601.21164","citing_title":"Concise Geometric Description as a Bridge: Unleashing the Potential of LLM for Plane Geometry Problem Solving","ref_index":10,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CAH45CJZEXVLNYYSISZLVA53PL","json":"https://pith.science/pith/CAH45CJZEXVLNYYSISZLVA53PL.json","graph_json":"https://pith.science/api/pith-number/CAH45CJZEXVLNYYSISZLVA53PL/graph.json","events_json":"https://pith.science/api/pith-number/CAH45CJZEXVLNYYSISZLVA53PL/events.json","paper":"https://pith.science/paper/CAH45CJZ"},"agent_actions":{"view_html":"https://pith.science/pith/CAH45CJZEXVLNYYSISZLVA53PL","download_json":"https://pith.science/pith/CAH45CJZEXVLNYYSISZLVA53PL.json","view_paper":"https://pith.science/paper/CAH45CJZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.17609&json=true","fetch_graph":"https://pith.science/api/pith-number/CAH45CJZEXVLNYYSISZLVA53PL/graph.json","fetch_events":"https://pith.science/api/pith-number/CAH45CJZEXVLNYYSISZLVA53PL/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CAH45CJZEXVLNYYSISZLVA53PL/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CAH45CJZEXVLNYYSISZLVA53PL/action/storage_attestation","attest_author":"https://pith.science/pith/CAH45CJZEXVLNYYSISZLVA53PL/action/author_attestation","sign_citation":"https://pith.science/pith/CAH45CJZEXVLNYYSISZLVA53PL/action/citation_signature","submit_replication":"https://pith.science/pith/CAH45CJZEXVLNYYSISZLVA53PL/action/replication_record"}},"created_at":"2026-07-05T11:53:03.368392+00:00","updated_at":"2026-07-05T11:53:03.368392+00:00"}