{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:RR35HHR3SSR64JGOS7543GP5Z2","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"9ea870095e492d672e296c3d73d3d11d9d3e9b71d8ef8ca681e4847ef4bb619d","cross_cats_sorted":["cs.SE"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-13T09:19:39Z","title_canon_sha256":"bfa47f7ff6d8fcedf058a50d520f791ff9308234d1a2f2df080e78e9d2f57b41"},"schema_version":"1.0","source":{"id":"2605.13229","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2605.13229","created_at":"2026-05-18T02:44:49Z"},{"alias_kind":"arxiv_version","alias_value":"2605.13229v1","created_at":"2026-05-18T02:44:49Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.13229","created_at":"2026-05-18T02:44:49Z"},{"alias_kind":"pith_short_12","alias_value":"RR35HHR3SSR6","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_16","alias_value":"RR35HHR3SSR64JGO","created_at":"2026-05-18T12:33:37Z"},{"alias_kind":"pith_short_8","alias_value":"RR35HHR3","created_at":"2026-05-18T12:33:37Z"}],"graph_snapshots":[{"event_id":"sha256:898f50ce96c1bf4cb416ba9e523d41f8f800a4bb4e71e103ad47a40dc5f4f14c","target":"graph","created_at":"2026-05-18T02:44:49Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":4,"items":[{"attestation":"unclaimed","claim_id":"C1","kind":"strongest_claim","source":"verdict.strongest_claim","status":"machine_extracted","text":"Through contrastive learning, we train a cross-lingual semantic model to directly assess functional equivalence between source and translated code. By formulating code translation as a multi-objective optimization problem, this robust semantic signal is seamlessly unified with compiler-based syntactic feedback within the direct preference optimization framework."},{"attestation":"unclaimed","claim_id":"C2","kind":"weakest_assumption","source":"verdict.weakest_assumption","status":"machine_extracted","text":"A robust semantic reward for code translation must be derived directly from the source code via a contrastively trained cross-lingual model that accurately captures functional equivalence without test cases or reference translations."},{"attestation":"unclaimed","claim_id":"C3","kind":"one_line_summary","source":"verdict.one_line_summary","status":"machine_extracted","text":"CTO improves code translation by training a semantic equivalence model through contrastive learning and unifying it with syntactic compiler feedback in a multi-objective direct preference optimization setup."},{"attestation":"unclaimed","claim_id":"C4","kind":"headline","source":"verdict.pith_extraction.headline","status":"machine_extracted","text":"A contrastively trained cross-lingual model supplies reliable semantic rewards for code translation inside direct preference optimization."}],"snapshot_sha256":"365d5d858247af58db1458fe1cdfbcd127aa4cd837f9f8ac9d9941a1e031c9db"},"formal_canon":{"evidence_count":2,"snapshot_sha256":"37477977d974d20a12970ce0f8c3bde56d5b88f0422ca8dfda9c07460404a405"},"paper":{"abstract_excerpt":"LLMs have shown immense potential for code translation, yet they often struggle to ensure both syntactic correctness and semantic consistency. While preference-based learning offers a promising alignment strategy, it is hindered by unreliable semantic rewards derived from sparse test cases or restrictive reference translations. We argue that a robust semantic reward for code translation must be derived directly from the source code. In this paper, we propose CTO to improve code translation with syntax-guided and semantic-aware preference optimization. Through contrastive learning, we train a c","authors_text":"Chen Shen, Huan Zhang, Jingyue Yang, Wei Cheng, Wei Hu, Yuhan Wu","cross_cats":["cs.SE"],"headline":"A contrastively trained cross-lingual model supplies reliable semantic rewards for code translation inside direct preference optimization.","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-13T09:19:39Z","title":"Improving Code Translation with Syntax-Guided and Semantic-aware Preference Optimization"},"references":{"count":50,"internal_anchors":0,"resolved_work":50,"sample":[{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":1,"title":"NeurIPS , year = 2020, pages =","work_id":"44b7714c-76ec-4291-940b-333e55cef989","year":2020},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":2,"title":"ICSE , year = 2024, pages =","work_id":"5af42412-6224-44e6-b327-27b2d40d0fe0","year":2024},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":3,"title":"ICLR , year = 2022, pages =","work_id":"cecdd246-be52-4da6-a91a-70ba65f4383d","year":2022},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":4,"title":"ICLR , year = 2023, pages =","work_id":"aec077c7-0b75-4a0a-83c4-509d3ca7249b","year":2023},{"cited_arxiv_id":"","doi":"","is_internal_anchor":false,"ref_index":5,"title":"ASE , year = 2023, pages =","work_id":"81bc644d-2da1-4838-ba74-926922b727b7","year":2023}],"snapshot_sha256":"f2b2d95fa794615f4b6f3e4a577920259348bb1553aa5c16c1f07bf0c61e0cc6"},"source":{"id":"2605.13229","kind":"arxiv","version":1},"verdict":{"created_at":"2026-05-14T20:00:12.450098Z","id":"1d41bc17-d481-404b-bfa7-28b4b86ad7ee","model_set":{"reader":"grok-4.3"},"one_line_summary":"CTO improves code translation by training a semantic equivalence model through contrastive learning and unifying it with syntactic compiler feedback in a multi-objective direct preference optimization setup.","pipeline_version":"pith-pipeline@v0.9.0","pith_extraction_headline":"A contrastively trained cross-lingual model supplies reliable semantic rewards for code translation inside direct preference optimization.","strongest_claim":"Through contrastive learning, we train a cross-lingual semantic model to directly assess functional equivalence between source and translated code. By formulating code translation as a multi-objective optimization problem, this robust semantic signal is seamlessly unified with compiler-based syntactic feedback within the direct preference optimization framework.","weakest_assumption":"A robust semantic reward for code translation must be derived directly from the source code via a contrastively trained cross-lingual model that accurately captures functional equivalence without test cases or reference translations."}},"verdict_id":"1d41bc17-d481-404b-bfa7-28b4b86ad7ee"}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:52c38456a10a2577d6df995b573d72cc5fb37097d3c24c951ae4c38605939d21","target":"record","created_at":"2026-05-18T02:44:49Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"9ea870095e492d672e296c3d73d3d11d9d3e9b71d8ef8ca681e4847ef4bb619d","cross_cats_sorted":["cs.SE"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.AI","submitted_at":"2026-05-13T09:19:39Z","title_canon_sha256":"bfa47f7ff6d8fcedf058a50d520f791ff9308234d1a2f2df080e78e9d2f57b41"},"schema_version":"1.0","source":{"id":"2605.13229","kind":"arxiv","version":1}},"canonical_sha256":"8c77d39e3b94a3ee24ce97fbcd99fdce9cf33e25db9facab778c90e059c6555b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"8c77d39e3b94a3ee24ce97fbcd99fdce9cf33e25db9facab778c90e059c6555b","first_computed_at":"2026-05-18T02:44:49.603180Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T02:44:49.603180Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"CNpq7bgqYy6m9+v5HaubBLLZcKY9t4b9MAFqDYJTc7t0aAKHSjV3stsu1dy0zB9Kwxf26MrP3weUjRTKOV2HDw==","signature_status":"signed_v1","signed_at":"2026-05-18T02:44:49.603674Z","signed_message":"canonical_sha256_bytes"},"source_id":"2605.13229","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:52c38456a10a2577d6df995b573d72cc5fb37097d3c24c951ae4c38605939d21","sha256:898f50ce96c1bf4cb416ba9e523d41f8f800a4bb4e71e103ad47a40dc5f4f14c"],"state_sha256":"5aaca21976150072d7cc05a8c64c1d66d80ce110f4ba9d2d6b08261c89f37d3b"}