{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:EVBCKCFKZF4MOIFA4HI2T6SUID","short_pith_number":"pith:EVBCKCFK","schema_version":"1.0","canonical_sha256":"25422508aac978c720a0e1d1a9fa5440e29080ace31fea60261b821c909cabae","source":{"kind":"arxiv","id":"2506.01277","version":1},"attestation_state":"computed","paper":{"title":"GeoLocSFT: Efficient Visual Geolocation via Supervised Fine-Tuning of Multimodal Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Lianlei Shan, Qiang Yi","submitted_at":"2025-06-02T03:16:19Z","abstract_excerpt":"Accurately determining the geographic location where a single image was taken, visual geolocation, remains a formidable challenge due to the planet's vastness and the deceptive similarity among distant locations. We introduce GeoLocSFT, a framework that demonstrates how targeted supervised fine-tuning (SFT) of a large multimodal foundation model (Gemma 3) using a small, high-quality dataset can yield highly competitive geolocation performance. GeoLocSFT is trained with only 2700 carefully selected image-GPS pairs from our geographically diverse MR600k dataset. Despite this limited data, our SF"},"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":"2506.01277","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2025-06-02T03:16:19Z","cross_cats_sorted":[],"title_canon_sha256":"2038bd35e77cb6a759b3482465c613fe7935974b6d27df32527cbb5605e18079","abstract_canon_sha256":"ef96ae35c07272ffb17157a73c8602893d8ff1e755ee9e5bda6d8c5bde6f3a3e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:14:11.120718Z","signature_b64":"FhDGjL7asY0dGapyfsl25Rt1Fy58ZZqw9uEwf0cGJpr6ShB252Ylcio2Usn3o4a4Mxg5mfvtJomfa9X9C+12CQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"25422508aac978c720a0e1d1a9fa5440e29080ace31fea60261b821c909cabae","last_reissued_at":"2026-07-05T11:14:11.120256Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:14:11.120256Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"GeoLocSFT: Efficient Visual Geolocation via Supervised Fine-Tuning of Multimodal Foundation Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Lianlei Shan, Qiang Yi","submitted_at":"2025-06-02T03:16:19Z","abstract_excerpt":"Accurately determining the geographic location where a single image was taken, visual geolocation, remains a formidable challenge due to the planet's vastness and the deceptive similarity among distant locations. We introduce GeoLocSFT, a framework that demonstrates how targeted supervised fine-tuning (SFT) of a large multimodal foundation model (Gemma 3) using a small, high-quality dataset can yield highly competitive geolocation performance. GeoLocSFT is trained with only 2700 carefully selected image-GPS pairs from our geographically diverse MR600k dataset. Despite this limited data, our SF"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.01277","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/2506.01277/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":"2506.01277","created_at":"2026-07-05T11:14:11.120312+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.01277v1","created_at":"2026-07-05T11:14:11.120312+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.01277","created_at":"2026-07-05T11:14:11.120312+00:00"},{"alias_kind":"pith_short_12","alias_value":"EVBCKCFKZF4M","created_at":"2026-07-05T11:14:11.120312+00:00"},{"alias_kind":"pith_short_16","alias_value":"EVBCKCFKZF4MOIFA","created_at":"2026-07-05T11:14:11.120312+00:00"},{"alias_kind":"pith_short_8","alias_value":"EVBCKCFK","created_at":"2026-07-05T11:14:11.120312+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.07642","citing_title":"Do VLMs See What Sensors Feel? A Scalable Expert-Guided Design for Wheelchair Accessibility Assessment from Street View","ref_index":37,"is_internal_anchor":false},{"citing_arxiv_id":"2508.01608","citing_title":"From Pixels to Places: A Systematic Benchmark for Evaluating Image Geolocalization Ability in Large Language Models","ref_index":47,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/EVBCKCFKZF4MOIFA4HI2T6SUID","json":"https://pith.science/pith/EVBCKCFKZF4MOIFA4HI2T6SUID.json","graph_json":"https://pith.science/api/pith-number/EVBCKCFKZF4MOIFA4HI2T6SUID/graph.json","events_json":"https://pith.science/api/pith-number/EVBCKCFKZF4MOIFA4HI2T6SUID/events.json","paper":"https://pith.science/paper/EVBCKCFK"},"agent_actions":{"view_html":"https://pith.science/pith/EVBCKCFKZF4MOIFA4HI2T6SUID","download_json":"https://pith.science/pith/EVBCKCFKZF4MOIFA4HI2T6SUID.json","view_paper":"https://pith.science/paper/EVBCKCFK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.01277&json=true","fetch_graph":"https://pith.science/api/pith-number/EVBCKCFKZF4MOIFA4HI2T6SUID/graph.json","fetch_events":"https://pith.science/api/pith-number/EVBCKCFKZF4MOIFA4HI2T6SUID/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/EVBCKCFKZF4MOIFA4HI2T6SUID/action/timestamp_anchor","attest_storage":"https://pith.science/pith/EVBCKCFKZF4MOIFA4HI2T6SUID/action/storage_attestation","attest_author":"https://pith.science/pith/EVBCKCFKZF4MOIFA4HI2T6SUID/action/author_attestation","sign_citation":"https://pith.science/pith/EVBCKCFKZF4MOIFA4HI2T6SUID/action/citation_signature","submit_replication":"https://pith.science/pith/EVBCKCFKZF4MOIFA4HI2T6SUID/action/replication_record"}},"created_at":"2026-07-05T11:14:11.120312+00:00","updated_at":"2026-07-05T11:14:11.120312+00:00"}