{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:UM43MBLI2P7JM73LC4UYJEYZCT","short_pith_number":"pith:UM43MBLI","schema_version":"1.0","canonical_sha256":"a339b60568d3fe967f6b172984931914c837879348838306843a05d36da3fc4a","source":{"kind":"arxiv","id":"2607.20501","version":1},"attestation_state":"computed","paper":{"title":"MKEvolve: A Modular Multi-Agent Framework for Kernel Code Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.MA"],"primary_cat":"cs.AI","authors_text":"Jason Yoo, Rajarshi Saha, Shaowei Zhu, Tao Yu, Wei Tang, Youngsuk Park","submitted_at":"2026-06-20T23:14:07Z","abstract_excerpt":"Despite rapid progress in LLM-based code generation, writing correct and performant kernels for hardware accelerators remains a key bottleneck in scaling modern ML workloads. We present MKEvolve (Modular Kernel Evolve), a framework that iteratively co-evolves a modular decomposition of complex PyTorch modules and the LLM-generated kernel for each submodule, refining the decomposition by splitting and fusing across iterations while independently improving each subkernel via LLM-driven beam search. The resulting kernels are programmatic compositions of independently verified subkernels, making t"},"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":"2607.20501","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2026-06-20T23:14:07Z","cross_cats_sorted":["cs.MA"],"title_canon_sha256":"f6e7db57c55693a4c71030c83d282a0a095a33feb1227d22cb0f60e30f3a7634","abstract_canon_sha256":"3dec2af77df051b7ca323e90a0ce62007eebace72ab90d8b48cf6dbf662cda20"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-24T00:23:19.194568Z","signature_b64":"mTnwFk4eEHfwc6ZjQDtu9DKA+jIcr7it1L0RTIUZqFNQkcXOYA8M4i8KpG3a7SQsdhrXitp/Yusvbdt38LtEBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a339b60568d3fe967f6b172984931914c837879348838306843a05d36da3fc4a","last_reissued_at":"2026-07-24T00:23:19.193698Z","signature_status":"signed_v1","first_computed_at":"2026-07-24T00:23:19.193698Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MKEvolve: A Modular Multi-Agent Framework for Kernel Code Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.MA"],"primary_cat":"cs.AI","authors_text":"Jason Yoo, Rajarshi Saha, Shaowei Zhu, Tao Yu, Wei Tang, Youngsuk Park","submitted_at":"2026-06-20T23:14:07Z","abstract_excerpt":"Despite rapid progress in LLM-based code generation, writing correct and performant kernels for hardware accelerators remains a key bottleneck in scaling modern ML workloads. We present MKEvolve (Modular Kernel Evolve), a framework that iteratively co-evolves a modular decomposition of complex PyTorch modules and the LLM-generated kernel for each submodule, refining the decomposition by splitting and fusing across iterations while independently improving each subkernel via LLM-driven beam search. The resulting kernels are programmatic compositions of independently verified subkernels, making t"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2607.20501","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/2607.20501/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":"2607.20501","created_at":"2026-07-24T00:23:19.194145+00:00"},{"alias_kind":"arxiv_version","alias_value":"2607.20501v1","created_at":"2026-07-24T00:23:19.194145+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2607.20501","created_at":"2026-07-24T00:23:19.194145+00:00"},{"alias_kind":"pith_short_12","alias_value":"UM43MBLI2P7J","created_at":"2026-07-24T00:23:19.194145+00:00"},{"alias_kind":"pith_short_16","alias_value":"UM43MBLI2P7JM73L","created_at":"2026-07-24T00:23:19.194145+00:00"},{"alias_kind":"pith_short_8","alias_value":"UM43MBLI","created_at":"2026-07-24T00:23:19.194145+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/UM43MBLI2P7JM73LC4UYJEYZCT","json":"https://pith.science/pith/UM43MBLI2P7JM73LC4UYJEYZCT.json","graph_json":"https://pith.science/api/pith-number/UM43MBLI2P7JM73LC4UYJEYZCT/graph.json","events_json":"https://pith.science/api/pith-number/UM43MBLI2P7JM73LC4UYJEYZCT/events.json","paper":"https://pith.science/paper/UM43MBLI"},"agent_actions":{"view_html":"https://pith.science/pith/UM43MBLI2P7JM73LC4UYJEYZCT","download_json":"https://pith.science/pith/UM43MBLI2P7JM73LC4UYJEYZCT.json","view_paper":"https://pith.science/paper/UM43MBLI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2607.20501&json=true","fetch_graph":"https://pith.science/api/pith-number/UM43MBLI2P7JM73LC4UYJEYZCT/graph.json","fetch_events":"https://pith.science/api/pith-number/UM43MBLI2P7JM73LC4UYJEYZCT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UM43MBLI2P7JM73LC4UYJEYZCT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UM43MBLI2P7JM73LC4UYJEYZCT/action/storage_attestation","attest_author":"https://pith.science/pith/UM43MBLI2P7JM73LC4UYJEYZCT/action/author_attestation","sign_citation":"https://pith.science/pith/UM43MBLI2P7JM73LC4UYJEYZCT/action/citation_signature","submit_replication":"https://pith.science/pith/UM43MBLI2P7JM73LC4UYJEYZCT/action/replication_record"}},"created_at":"2026-07-24T00:23:19.194145+00:00","updated_at":"2026-07-24T00:23:19.194145+00:00"}