{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:LSKTSRBASNBJ44QDMOQKZFU7F3","short_pith_number":"pith:LSKTSRBA","schema_version":"1.0","canonical_sha256":"5c9539442093429e720363a0ac969f2ec69588ec78863d389a92a111fe07f735","source":{"kind":"arxiv","id":"2504.10187","version":2},"attestation_state":"computed","paper":{"title":"DeepTrans: Deep Reasoning Translation via Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Fandong Meng, Jiaan Wang, Jie Zhou","submitted_at":"2025-04-14T12:40:39Z","abstract_excerpt":"Recently, deep reasoning LLMs (e.g., OpenAI o1 and DeepSeek-R1) have shown promising performance in various downstream tasks. Free translation is an important and interesting task in the multilingual world, which requires going beyond word-for-word translation. However, the task is still under-explored in deep reasoning LLMs. In this paper, we introduce DeepTrans, a deep reasoning translation model that learns free translation via reinforcement learning (RL). Specifically, we carefully build a reward model with pre-defined scoring criteria on both the translation results and the thought proces"},"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":"2504.10187","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-04-14T12:40:39Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ffc1105890f928149bc5fa61ed74c9e8e4a0732fbd5c0ffd44931eaf16b0eeb4","abstract_canon_sha256":"1e604460e8382bd84be2712ed12e3ad7c4db373c92353c9eb8557550c7853445"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:01:21.932175Z","signature_b64":"Llcu1zdwuhBZrEd2HpDg0L7HlrOu45jXKgrat+MThGuTc/cyXx4nY7aoIMaRWRnR0/FVKv3LIGPrpRUeJib4CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"5c9539442093429e720363a0ac969f2ec69588ec78863d389a92a111fe07f735","last_reissued_at":"2026-07-05T12:01:21.931505Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:01:21.931505Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DeepTrans: Deep Reasoning Translation via Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Fandong Meng, Jiaan Wang, Jie Zhou","submitted_at":"2025-04-14T12:40:39Z","abstract_excerpt":"Recently, deep reasoning LLMs (e.g., OpenAI o1 and DeepSeek-R1) have shown promising performance in various downstream tasks. Free translation is an important and interesting task in the multilingual world, which requires going beyond word-for-word translation. However, the task is still under-explored in deep reasoning LLMs. In this paper, we introduce DeepTrans, a deep reasoning translation model that learns free translation via reinforcement learning (RL). Specifically, we carefully build a reward model with pre-defined scoring criteria on both the translation results and the thought proces"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2504.10187","kind":"arxiv","version":2},"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/2504.10187/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":"2504.10187","created_at":"2026-07-05T12:01:21.931592+00:00"},{"alias_kind":"arxiv_version","alias_value":"2504.10187v2","created_at":"2026-07-05T12:01:21.931592+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2504.10187","created_at":"2026-07-05T12:01:21.931592+00:00"},{"alias_kind":"pith_short_12","alias_value":"LSKTSRBASNBJ","created_at":"2026-07-05T12:01:21.931592+00:00"},{"alias_kind":"pith_short_16","alias_value":"LSKTSRBASNBJ44QD","created_at":"2026-07-05T12:01:21.931592+00:00"},{"alias_kind":"pith_short_8","alias_value":"LSKTSRBA","created_at":"2026-07-05T12:01:21.931592+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.29481","citing_title":"To Reason or to Fabricate: Reasoning Without Shortcuts via Hint-Anchored Pairwise Aggregation","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2604.18490","citing_title":"LQM: Linguistically Motivated Multidimensional Quality Metrics for Machine Translation","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19144","citing_title":"ReflectMT: Internalizing Reflection for Efficient and High-Quality Machine Translation","ref_index":49,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/LSKTSRBASNBJ44QDMOQKZFU7F3","json":"https://pith.science/pith/LSKTSRBASNBJ44QDMOQKZFU7F3.json","graph_json":"https://pith.science/api/pith-number/LSKTSRBASNBJ44QDMOQKZFU7F3/graph.json","events_json":"https://pith.science/api/pith-number/LSKTSRBASNBJ44QDMOQKZFU7F3/events.json","paper":"https://pith.science/paper/LSKTSRBA"},"agent_actions":{"view_html":"https://pith.science/pith/LSKTSRBASNBJ44QDMOQKZFU7F3","download_json":"https://pith.science/pith/LSKTSRBASNBJ44QDMOQKZFU7F3.json","view_paper":"https://pith.science/paper/LSKTSRBA","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2504.10187&json=true","fetch_graph":"https://pith.science/api/pith-number/LSKTSRBASNBJ44QDMOQKZFU7F3/graph.json","fetch_events":"https://pith.science/api/pith-number/LSKTSRBASNBJ44QDMOQKZFU7F3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/LSKTSRBASNBJ44QDMOQKZFU7F3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/LSKTSRBASNBJ44QDMOQKZFU7F3/action/storage_attestation","attest_author":"https://pith.science/pith/LSKTSRBASNBJ44QDMOQKZFU7F3/action/author_attestation","sign_citation":"https://pith.science/pith/LSKTSRBASNBJ44QDMOQKZFU7F3/action/citation_signature","submit_replication":"https://pith.science/pith/LSKTSRBASNBJ44QDMOQKZFU7F3/action/replication_record"}},"created_at":"2026-07-05T12:01:21.931592+00:00","updated_at":"2026-07-05T12:01:21.931592+00:00"}