{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:GC4LIP5NH66NKL4BA4RTHUPIAC","short_pith_number":"pith:GC4LIP5N","schema_version":"1.0","canonical_sha256":"30b8b43fad3fbcd52f81072333d1e8008afac58d9b9d374e984f9a1e84e7bafd","source":{"kind":"arxiv","id":"2510.14331","version":3},"attestation_state":"computed","paper":{"title":"LLM Priors for ERM over Programs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Eran Malach, Priyadarsi Mishra, Shivam Singhal, Tomer Galanti","submitted_at":"2025-10-16T06:10:11Z","abstract_excerpt":"We study program-learning methods that are efficient in both samples and computation. Classical learning theory suggests that when the target admits a short program description, for example a short piece of ``Python code'', it can be learned from few examples by ERM over the program class. However, this approach relies on enumerating candidate programs, which is typically exponential in the description length; gradient-based training avoids this explicit search but, for some families of short programs, can require exponentially many samples to succeed. We propose \\textsc{LLM-PV}, a propose-and"},"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":"2510.14331","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2025-10-16T06:10:11Z","cross_cats_sorted":[],"title_canon_sha256":"e2f4f6a346271267c5ef888849d66f0b65e2cc0f0e4fcf6c48d7636343384b45","abstract_canon_sha256":"88f3f119194e350de8ad4ac9c6c7b5db2403897c60d916e2b5bbc6a3ae3a1f84"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-03T01:17:13.718699Z","signature_b64":"j8NDvpdhEQa1M2sPLFgyWFzTpSQ/EBSFLTrgkviZMnhhpdLwQG2i1/KWMe7AO8uwg/hJMhVkF8wlP1JP71aaDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"30b8b43fad3fbcd52f81072333d1e8008afac58d9b9d374e984f9a1e84e7bafd","last_reissued_at":"2026-07-03T01:17:13.718174Z","signature_status":"signed_v1","first_computed_at":"2026-07-03T01:17:13.718174Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LLM Priors for ERM over Programs","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Eran Malach, Priyadarsi Mishra, Shivam Singhal, Tomer Galanti","submitted_at":"2025-10-16T06:10:11Z","abstract_excerpt":"We study program-learning methods that are efficient in both samples and computation. Classical learning theory suggests that when the target admits a short program description, for example a short piece of ``Python code'', it can be learned from few examples by ERM over the program class. However, this approach relies on enumerating candidate programs, which is typically exponential in the description length; gradient-based training avoids this explicit search but, for some families of short programs, can require exponentially many samples to succeed. We propose \\textsc{LLM-PV}, a propose-and"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2510.14331","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/2510.14331/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":"2510.14331","created_at":"2026-07-03T01:17:13.718225+00:00"},{"alias_kind":"arxiv_version","alias_value":"2510.14331v3","created_at":"2026-07-03T01:17:13.718225+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2510.14331","created_at":"2026-07-03T01:17:13.718225+00:00"},{"alias_kind":"pith_short_12","alias_value":"GC4LIP5NH66N","created_at":"2026-07-03T01:17:13.718225+00:00"},{"alias_kind":"pith_short_16","alias_value":"GC4LIP5NH66NKL4B","created_at":"2026-07-03T01:17:13.718225+00:00"},{"alias_kind":"pith_short_8","alias_value":"GC4LIP5N","created_at":"2026-07-03T01:17:13.718225+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2605.20086","citing_title":"What Do Evolutionary Coding Agents Evolve?","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2605.05485","citing_title":"ReaComp: Compiling LLM Reasoning into Symbolic Solvers for Efficient Program Synthesis","ref_index":8,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GC4LIP5NH66NKL4BA4RTHUPIAC","json":"https://pith.science/pith/GC4LIP5NH66NKL4BA4RTHUPIAC.json","graph_json":"https://pith.science/api/pith-number/GC4LIP5NH66NKL4BA4RTHUPIAC/graph.json","events_json":"https://pith.science/api/pith-number/GC4LIP5NH66NKL4BA4RTHUPIAC/events.json","paper":"https://pith.science/paper/GC4LIP5N"},"agent_actions":{"view_html":"https://pith.science/pith/GC4LIP5NH66NKL4BA4RTHUPIAC","download_json":"https://pith.science/pith/GC4LIP5NH66NKL4BA4RTHUPIAC.json","view_paper":"https://pith.science/paper/GC4LIP5N","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2510.14331&json=true","fetch_graph":"https://pith.science/api/pith-number/GC4LIP5NH66NKL4BA4RTHUPIAC/graph.json","fetch_events":"https://pith.science/api/pith-number/GC4LIP5NH66NKL4BA4RTHUPIAC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GC4LIP5NH66NKL4BA4RTHUPIAC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GC4LIP5NH66NKL4BA4RTHUPIAC/action/storage_attestation","attest_author":"https://pith.science/pith/GC4LIP5NH66NKL4BA4RTHUPIAC/action/author_attestation","sign_citation":"https://pith.science/pith/GC4LIP5NH66NKL4BA4RTHUPIAC/action/citation_signature","submit_replication":"https://pith.science/pith/GC4LIP5NH66NKL4BA4RTHUPIAC/action/replication_record"}},"created_at":"2026-07-03T01:17:13.718225+00:00","updated_at":"2026-07-03T01:17:13.718225+00:00"}