{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:UMIYPX4TOZ6D2GBI26ZO563HLB","short_pith_number":"pith:UMIYPX4T","schema_version":"1.0","canonical_sha256":"a31187df93767c3d1828d7b2eefb67587cb19d9434dbb31764522bf73ff6cbee","source":{"kind":"arxiv","id":"2106.04985","version":1},"attestation_state":"computed","paper":{"title":"Energy-Based Models for Code Generation under Compilability Constraints","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.NE","cs.SE"],"primary_cat":"cs.LG","authors_text":"Germ\\'an Kruszewski, Hady Elsahar, Marc Dymetman, Tomasz Korbak","submitted_at":"2021-06-09T11:06:32Z","abstract_excerpt":"Neural language models can be successfully trained on source code, leading to applications such as code completion. However, their versatile autoregressive self-supervision objective overlooks important global sequence-level features that are present in the data such as syntactic correctness or compilability. In this work, we pose the problem of learning to generate compilable code as constraint satisfaction. We define an Energy-Based Model (EBM) representing a pre-trained generative model with an imposed constraint of generating only compilable sequences. We then use the KL-Adaptive Distribut"},"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":"2106.04985","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-06-09T11:06:32Z","cross_cats_sorted":["cs.CL","cs.NE","cs.SE"],"title_canon_sha256":"1b3018147bec15ea4dbb131b6d22a348a672c6a97aef21370cafaa0c9ab34d9e","abstract_canon_sha256":"10d7414bd7564e3612779ce49f486423be078db8bb0d725daf937b546ba30dd8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:47:51.749026Z","signature_b64":"XhwYIKocu5nyoM4n/6qzESpW30nVXphUGyaNsVcfYi7EJHYeF9xM03muGiLLIdXdlXboZNV6HFFfSxYwIsC7DQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a31187df93767c3d1828d7b2eefb67587cb19d9434dbb31764522bf73ff6cbee","last_reissued_at":"2026-07-05T02:47:51.748586Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:47:51.748586Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Energy-Based Models for Code Generation under Compilability Constraints","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL","cs.NE","cs.SE"],"primary_cat":"cs.LG","authors_text":"Germ\\'an Kruszewski, Hady Elsahar, Marc Dymetman, Tomasz Korbak","submitted_at":"2021-06-09T11:06:32Z","abstract_excerpt":"Neural language models can be successfully trained on source code, leading to applications such as code completion. However, their versatile autoregressive self-supervision objective overlooks important global sequence-level features that are present in the data such as syntactic correctness or compilability. In this work, we pose the problem of learning to generate compilable code as constraint satisfaction. We define an Energy-Based Model (EBM) representing a pre-trained generative model with an imposed constraint of generating only compilable sequences. We then use the KL-Adaptive Distribut"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.04985","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/2106.04985/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":"2106.04985","created_at":"2026-07-05T02:47:51.748643+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.04985v1","created_at":"2026-07-05T02:47:51.748643+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.04985","created_at":"2026-07-05T02:47:51.748643+00:00"},{"alias_kind":"pith_short_12","alias_value":"UMIYPX4TOZ6D","created_at":"2026-07-05T02:47:51.748643+00:00"},{"alias_kind":"pith_short_16","alias_value":"UMIYPX4TOZ6D2GBI","created_at":"2026-07-05T02:47:51.748643+00:00"},{"alias_kind":"pith_short_8","alias_value":"UMIYPX4T","created_at":"2026-07-05T02:47:51.748643+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2505.01111","citing_title":"Incorporating Inductive Biases to Energy-based Generative Models","ref_index":2010,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UMIYPX4TOZ6D2GBI26ZO563HLB","json":"https://pith.science/pith/UMIYPX4TOZ6D2GBI26ZO563HLB.json","graph_json":"https://pith.science/api/pith-number/UMIYPX4TOZ6D2GBI26ZO563HLB/graph.json","events_json":"https://pith.science/api/pith-number/UMIYPX4TOZ6D2GBI26ZO563HLB/events.json","paper":"https://pith.science/paper/UMIYPX4T"},"agent_actions":{"view_html":"https://pith.science/pith/UMIYPX4TOZ6D2GBI26ZO563HLB","download_json":"https://pith.science/pith/UMIYPX4TOZ6D2GBI26ZO563HLB.json","view_paper":"https://pith.science/paper/UMIYPX4T","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.04985&json=true","fetch_graph":"https://pith.science/api/pith-number/UMIYPX4TOZ6D2GBI26ZO563HLB/graph.json","fetch_events":"https://pith.science/api/pith-number/UMIYPX4TOZ6D2GBI26ZO563HLB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UMIYPX4TOZ6D2GBI26ZO563HLB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UMIYPX4TOZ6D2GBI26ZO563HLB/action/storage_attestation","attest_author":"https://pith.science/pith/UMIYPX4TOZ6D2GBI26ZO563HLB/action/author_attestation","sign_citation":"https://pith.science/pith/UMIYPX4TOZ6D2GBI26ZO563HLB/action/citation_signature","submit_replication":"https://pith.science/pith/UMIYPX4TOZ6D2GBI26ZO563HLB/action/replication_record"}},"created_at":"2026-07-05T02:47:51.748643+00:00","updated_at":"2026-07-05T02:47:51.748643+00:00"}