{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:IEJG6HMDBMQUUMFZNOWJMTRON7","short_pith_number":"pith:IEJG6HMD","schema_version":"1.0","canonical_sha256":"41126f1d830b214a30b96bac964e2e6fc21c373f02cb6fdae263a20d142dbeac","source":{"kind":"arxiv","id":"2508.06492","version":1},"attestation_state":"computed","paper":{"title":"Effective Training Data Synthesis for Improving MLLM Chart Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Ali Payani, Gaowen Liu, Liang Zheng, Yuan-Sen Ting, Yunzhong Hou, Yuwei Yang, Zeyu Zhang, Zhuowan Li","submitted_at":"2025-08-08T17:59:10Z","abstract_excerpt":"Being able to effectively read scientific plots, or chart understanding, is a central part toward building effective agents for science. However, existing multimodal large language models (MLLMs), especially open-source ones, are still falling behind with a typical success rate of 30%-50% on challenging benchmarks. Previous studies on fine-tuning MLLMs with synthetic charts are often restricted by their inadequate similarity to the real charts, which could compromise model training and performance on complex real-world charts. In this study, we show that modularizing chart generation and diver"},"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":"2508.06492","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-08-08T17:59:10Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"e3d14b12184287bf0a9b81b7b151e62ebef1573758e8c4978af363c8582d6578","abstract_canon_sha256":"e76a3378d8c9bc39d49c26b797025e3f6d083634ed18d8de4218449851d1eb09"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:50:57.241078Z","signature_b64":"in+O99MkT6oetRikRsMKvwR8/cWpbOoXML+QHknUv1yFy2gucwTi+l5kvoETYtIPZAlYW0lOyviZqk9Z3T3sCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"41126f1d830b214a30b96bac964e2e6fc21c373f02cb6fdae263a20d142dbeac","last_reissued_at":"2026-07-05T11:50:57.240586Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:50:57.240586Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Effective Training Data Synthesis for Improving MLLM Chart Understanding","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Ali Payani, Gaowen Liu, Liang Zheng, Yuan-Sen Ting, Yunzhong Hou, Yuwei Yang, Zeyu Zhang, Zhuowan Li","submitted_at":"2025-08-08T17:59:10Z","abstract_excerpt":"Being able to effectively read scientific plots, or chart understanding, is a central part toward building effective agents for science. However, existing multimodal large language models (MLLMs), especially open-source ones, are still falling behind with a typical success rate of 30%-50% on challenging benchmarks. Previous studies on fine-tuning MLLMs with synthetic charts are often restricted by their inadequate similarity to the real charts, which could compromise model training and performance on complex real-world charts. In this study, we show that modularizing chart generation and diver"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.06492","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/2508.06492/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":"2508.06492","created_at":"2026-07-05T11:50:57.240646+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.06492v1","created_at":"2026-07-05T11:50:57.240646+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.06492","created_at":"2026-07-05T11:50:57.240646+00:00"},{"alias_kind":"pith_short_12","alias_value":"IEJG6HMDBMQU","created_at":"2026-07-05T11:50:57.240646+00:00"},{"alias_kind":"pith_short_16","alias_value":"IEJG6HMDBMQUUMFZ","created_at":"2026-07-05T11:50:57.240646+00:00"},{"alias_kind":"pith_short_8","alias_value":"IEJG6HMD","created_at":"2026-07-05T11:50:57.240646+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.07415","citing_title":"ChartREG++: Towards Benchmarking and Improving Chart Referring Expression Grounding under Diverse referring clues and Multi-Target Referring","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2601.13606","citing_title":"ChartVerse: Scaling Chart Reasoning via Reliable Programmatic Synthesis from Scratch","ref_index":45,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07415","citing_title":"ChartREG++: Towards Benchmarking and Improving Chart Referring Expression Grounding under Diverse referring clues and Multi-Target Referring","ref_index":24,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IEJG6HMDBMQUUMFZNOWJMTRON7","json":"https://pith.science/pith/IEJG6HMDBMQUUMFZNOWJMTRON7.json","graph_json":"https://pith.science/api/pith-number/IEJG6HMDBMQUUMFZNOWJMTRON7/graph.json","events_json":"https://pith.science/api/pith-number/IEJG6HMDBMQUUMFZNOWJMTRON7/events.json","paper":"https://pith.science/paper/IEJG6HMD"},"agent_actions":{"view_html":"https://pith.science/pith/IEJG6HMDBMQUUMFZNOWJMTRON7","download_json":"https://pith.science/pith/IEJG6HMDBMQUUMFZNOWJMTRON7.json","view_paper":"https://pith.science/paper/IEJG6HMD","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.06492&json=true","fetch_graph":"https://pith.science/api/pith-number/IEJG6HMDBMQUUMFZNOWJMTRON7/graph.json","fetch_events":"https://pith.science/api/pith-number/IEJG6HMDBMQUUMFZNOWJMTRON7/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IEJG6HMDBMQUUMFZNOWJMTRON7/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IEJG6HMDBMQUUMFZNOWJMTRON7/action/storage_attestation","attest_author":"https://pith.science/pith/IEJG6HMDBMQUUMFZNOWJMTRON7/action/author_attestation","sign_citation":"https://pith.science/pith/IEJG6HMDBMQUUMFZNOWJMTRON7/action/citation_signature","submit_replication":"https://pith.science/pith/IEJG6HMDBMQUUMFZNOWJMTRON7/action/replication_record"}},"created_at":"2026-07-05T11:50:57.240646+00:00","updated_at":"2026-07-05T11:50:57.240646+00:00"}