{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:SMN3TONEF2C4SQCZWXSEHZABRN","short_pith_number":"pith:SMN3TONE","schema_version":"1.0","canonical_sha256":"931bb9b9a42e85c94059b5e443e4018b5fc8cb1beb486eb9b6e0187759c44e0f","source":{"kind":"arxiv","id":"2506.10959","version":3},"attestation_state":"computed","paper":{"title":"Understanding In-Context Learning on Structured Manifolds: Bridging Attention to Kernel Methods","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","math.ST","stat.TH"],"primary_cat":"cs.LG","authors_text":"Alexander Hsu, Rongjie Lai, Wenjing Liao, Zhaiming Shen","submitted_at":"2025-06-12T17:56:26Z","abstract_excerpt":"While in-context learning (ICL) has achieved remarkable success in natural language and vision domains, its theoretical understanding-particularly in the context of structured geometric data-remains unexplored. This paper initiates a theoretical study of ICL for regression of H\\\"older functions on manifolds. We establish a novel connection between the attention mechanism and classical kernel methods, demonstrating that transformers effectively perform kernel-based prediction at a new query through its interaction with the prompt. This connection is validated by numerical experiments, revealing"},"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":"2506.10959","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2025-06-12T17:56:26Z","cross_cats_sorted":["cs.AI","math.ST","stat.TH"],"title_canon_sha256":"a14a85fa9854b8936ab5b83f9313d9079968072fb53ae0f95eef8fc0c78beb09","abstract_canon_sha256":"3693f6da28f0bfd714997055f5ede9e6fd3d6c2fa1aa4dc24b8cd14c43aaa01b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:02:51.105993Z","signature_b64":"4EE9bgTPHWhog5xY6jJwJKa2nnvJTdlCY9rpzwqg6UVmRurrViBA+IB4QD15oA+j5FiXasV5heXk0ZMUaCUCDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"931bb9b9a42e85c94059b5e443e4018b5fc8cb1beb486eb9b6e0187759c44e0f","last_reissued_at":"2026-05-20T00:02:51.105133Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:02:51.105133Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Understanding In-Context Learning on Structured Manifolds: Bridging Attention to Kernel Methods","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","math.ST","stat.TH"],"primary_cat":"cs.LG","authors_text":"Alexander Hsu, Rongjie Lai, Wenjing Liao, Zhaiming Shen","submitted_at":"2025-06-12T17:56:26Z","abstract_excerpt":"While in-context learning (ICL) has achieved remarkable success in natural language and vision domains, its theoretical understanding-particularly in the context of structured geometric data-remains unexplored. This paper initiates a theoretical study of ICL for regression of H\\\"older functions on manifolds. We establish a novel connection between the attention mechanism and classical kernel methods, demonstrating that transformers effectively perform kernel-based prediction at a new query through its interaction with the prompt. This connection is validated by numerical experiments, revealing"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.10959","kind":"arxiv","version":3},"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/2506.10959/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":"2506.10959","created_at":"2026-05-20T00:02:51.105266+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.10959v3","created_at":"2026-05-20T00:02:51.105266+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.10959","created_at":"2026-05-20T00:02:51.105266+00:00"},{"alias_kind":"pith_short_12","alias_value":"SMN3TONEF2C4","created_at":"2026-05-20T00:02:51.105266+00:00"},{"alias_kind":"pith_short_16","alias_value":"SMN3TONEF2C4SQCZ","created_at":"2026-05-20T00:02:51.105266+00:00"},{"alias_kind":"pith_short_8","alias_value":"SMN3TONE","created_at":"2026-05-20T00:02:51.105266+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":4,"internal_anchor_count":4,"sample":[{"citing_arxiv_id":"2505.03205","citing_title":"Transformers for Learning on Noisy and Task-Level Manifolds: Approximation and Generalization Insights","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08475","citing_title":"Transformers Can Implement Preconditioned Richardson Iteration for In-Context Gaussian Kernel Regression","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08475","citing_title":"Transformers Can Implement Preconditioned Richardson Iteration for In-Context Gaussian Kernel Regression","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2605.05176","citing_title":"Understanding In-Context Learning for Nonlinear Regression with Transformers: Attention as Featurizer","ref_index":32,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SMN3TONEF2C4SQCZWXSEHZABRN","json":"https://pith.science/pith/SMN3TONEF2C4SQCZWXSEHZABRN.json","graph_json":"https://pith.science/api/pith-number/SMN3TONEF2C4SQCZWXSEHZABRN/graph.json","events_json":"https://pith.science/api/pith-number/SMN3TONEF2C4SQCZWXSEHZABRN/events.json","paper":"https://pith.science/paper/SMN3TONE"},"agent_actions":{"view_html":"https://pith.science/pith/SMN3TONEF2C4SQCZWXSEHZABRN","download_json":"https://pith.science/pith/SMN3TONEF2C4SQCZWXSEHZABRN.json","view_paper":"https://pith.science/paper/SMN3TONE","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.10959&json=true","fetch_graph":"https://pith.science/api/pith-number/SMN3TONEF2C4SQCZWXSEHZABRN/graph.json","fetch_events":"https://pith.science/api/pith-number/SMN3TONEF2C4SQCZWXSEHZABRN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SMN3TONEF2C4SQCZWXSEHZABRN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SMN3TONEF2C4SQCZWXSEHZABRN/action/storage_attestation","attest_author":"https://pith.science/pith/SMN3TONEF2C4SQCZWXSEHZABRN/action/author_attestation","sign_citation":"https://pith.science/pith/SMN3TONEF2C4SQCZWXSEHZABRN/action/citation_signature","submit_replication":"https://pith.science/pith/SMN3TONEF2C4SQCZWXSEHZABRN/action/replication_record"}},"created_at":"2026-05-20T00:02:51.105266+00:00","updated_at":"2026-05-20T00:02:51.105266+00:00"}