{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:5HUVFJONZHHUAJGTN7GCEALNPM","short_pith_number":"pith:5HUVFJON","schema_version":"1.0","canonical_sha256":"e9e952a5cdc9cf4024d36fcc22016d7b096e3c9e11b7d7a667eed722d1b00223","source":{"kind":"arxiv","id":"2512.21414","version":2},"attestation_state":"computed","paper":{"title":"A Tool Bottleneck Framework for Clinically-Informed and Interpretable Medical Image Understanding","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Alan Q. Wang, Christina Liu, Ehsan Adeli, Jiajun Wu, Joy Hsu","submitted_at":"2025-12-24T20:30:01Z","abstract_excerpt":"Recent tool-use frameworks powered by vision-language models (VLMs) improve image understanding by grounding model predictions with specialized tools. Broadly, these frameworks leverage VLMs and a pre-specified toolbox to decompose the prediction task into multiple tool calls (often deep learning models) which are composed to make a prediction. The dominant approach to composing tools is using text, via function calls embedded in VLM-generated code or natural language. However, these methods often perform poorly on medical image understanding, where salient information is encoded as spatially-"},"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":"2512.21414","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CV","submitted_at":"2025-12-24T20:30:01Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"b6b1c7f0fc028b315e55749e3e574748472de8f23cf333be3aeaebf743ad0fa9","abstract_canon_sha256":"1f49b592d1ee5148b224c84da71b156480ded28fa830cafe078d83e4d99dc8d4"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-10T00:18:42.733689Z","signature_b64":"PRfmIoOtvczeKj3gHDM6UjLlvXHY6pOphnw1dhG5D8zw/LvJc+ZgcuEgoJnHMjXBXRPs93SV+fXLQTIA540oCw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e9e952a5cdc9cf4024d36fcc22016d7b096e3c9e11b7d7a667eed722d1b00223","last_reissued_at":"2026-07-10T00:18:42.733078Z","signature_status":"signed_v1","first_computed_at":"2026-07-10T00:18:42.733078Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Tool Bottleneck Framework for Clinically-Informed and Interpretable Medical Image Understanding","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"cs.CV","authors_text":"Alan Q. Wang, Christina Liu, Ehsan Adeli, Jiajun Wu, Joy Hsu","submitted_at":"2025-12-24T20:30:01Z","abstract_excerpt":"Recent tool-use frameworks powered by vision-language models (VLMs) improve image understanding by grounding model predictions with specialized tools. Broadly, these frameworks leverage VLMs and a pre-specified toolbox to decompose the prediction task into multiple tool calls (often deep learning models) which are composed to make a prediction. The dominant approach to composing tools is using text, via function calls embedded in VLM-generated code or natural language. However, these methods often perform poorly on medical image understanding, where salient information is encoded as spatially-"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2512.21414","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/2512.21414/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":"2512.21414","created_at":"2026-07-10T00:18:42.733148+00:00"},{"alias_kind":"arxiv_version","alias_value":"2512.21414v2","created_at":"2026-07-10T00:18:42.733148+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2512.21414","created_at":"2026-07-10T00:18:42.733148+00:00"},{"alias_kind":"pith_short_12","alias_value":"5HUVFJONZHHU","created_at":"2026-07-10T00:18:42.733148+00:00"},{"alias_kind":"pith_short_16","alias_value":"5HUVFJONZHHUAJGT","created_at":"2026-07-10T00:18:42.733148+00:00"},{"alias_kind":"pith_short_8","alias_value":"5HUVFJON","created_at":"2026-07-10T00:18:42.733148+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/5HUVFJONZHHUAJGTN7GCEALNPM","json":"https://pith.science/pith/5HUVFJONZHHUAJGTN7GCEALNPM.json","graph_json":"https://pith.science/api/pith-number/5HUVFJONZHHUAJGTN7GCEALNPM/graph.json","events_json":"https://pith.science/api/pith-number/5HUVFJONZHHUAJGTN7GCEALNPM/events.json","paper":"https://pith.science/paper/5HUVFJON"},"agent_actions":{"view_html":"https://pith.science/pith/5HUVFJONZHHUAJGTN7GCEALNPM","download_json":"https://pith.science/pith/5HUVFJONZHHUAJGTN7GCEALNPM.json","view_paper":"https://pith.science/paper/5HUVFJON","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2512.21414&json=true","fetch_graph":"https://pith.science/api/pith-number/5HUVFJONZHHUAJGTN7GCEALNPM/graph.json","fetch_events":"https://pith.science/api/pith-number/5HUVFJONZHHUAJGTN7GCEALNPM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5HUVFJONZHHUAJGTN7GCEALNPM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5HUVFJONZHHUAJGTN7GCEALNPM/action/storage_attestation","attest_author":"https://pith.science/pith/5HUVFJONZHHUAJGTN7GCEALNPM/action/author_attestation","sign_citation":"https://pith.science/pith/5HUVFJONZHHUAJGTN7GCEALNPM/action/citation_signature","submit_replication":"https://pith.science/pith/5HUVFJONZHHUAJGTN7GCEALNPM/action/replication_record"}},"created_at":"2026-07-10T00:18:42.733148+00:00","updated_at":"2026-07-10T00:18:42.733148+00:00"}