{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/alleviating-distribution-shift-in-synthetic","title":"Alleviating Distribution Shift in Synthetic Data for Machine Translation Quality Estimation","arxiv_id":"2502.19941","date":"2025-02-27","proceeding":null,"authors":["Xiang Geng","Zhejian Lai","Jiajun Chen","Hao Yang","ShuJian Huang"],"abstract":"Quality Estimation (QE) models evaluate the quality of machine translations without reference translations, serving as the reward models for the translation task. Due to the data scarcity, synthetic data generation has emerged as a promising solution. However, synthetic QE data often suffers from distribution shift, which can manifest as discrepancies between pseudo and real translations, or in pseudo labels that do not align with human preferences. To tackle this issue, we introduce ADSQE, a novel framework for alleviating distribution shift in synthetic QE data. To reduce the difference between pseudo and real translations, we employ the constrained beam search algorithm and enhance translation diversity through the use of distinct generation models. ADSQE uses references, i.e., translation supervision signals, to guide both the generation and annotation processes, enhancing the quality of word-level labels. ADSE further identifies the shortest phrase covering consecutive error tokens, mimicking human annotation behavior, to assign the final phrase-level labels. Specially, we underscore that the translation model can not annotate translations of itself accurately. Extensive experiments demonstrate that ADSQE outperforms SOTA baselines like COMET in both supervised and unsupervised settings. Further analysis offers insights into synthetic data generation that could benefit reward models for other tasks.","url_abs":"https://arxiv.org/abs/2502.19941v1","url_pdf":"https://arxiv.org/pdf/2502.19941v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[],"tasks":[{"task_slug":"machine-translation","task_name":"Machine Translation"},{"task_slug":"synthetic-data-generation","task_name":"Synthetic Data Generation"},{"task_slug":"translation","task_name":"Translation"}],"methods":[{"method_slug":"align","method_name":"ALIGN"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2502.19941","mcp":{"get_harvested_code_for_paper":{"arxiv_id":"2502.19941"}},"developers":"https://syntology.ai/developers","read_at":"2026-09-24T18:15:14+00:00","read_at_is":"when the build read Syntology's graph, not when any sample ran","claim":"Per-sample execution status on synthesized fixtures; not a correctness claim about the paper. Samples come from repositories linked to the paper, official or community; repo_kind says which.","repos":[{"provenance":"deterministic:regex_extraction","url":"https://github.com/NJUNLP/njuqe","reach":{"status":"ok","spdx":"MIT"}}],"summary":{"unverified":11},"by_repo_kind":{"found_in_text":{"samples":11,"ran":0,"repositories":1}},"repo_kind_vocabulary":{"official":"The archive marks this repository official for the paper","named_in_paper":"The archive records that the paper mentions this repository; it is not marked official","listed":"In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper","found_in_text":"Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted","community":"Not in the archive's code links for this paper; a community repository Syntology harvested"},"n_pointer_only_for_licence":0,"samples":[{"code_sha256_prefix":"e9324ea7d15a909c","entry":"args_update","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/modules/qe_detector.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/modules/qe_detector.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"e9324ea7d15a909c"}},{"code_sha256_prefix":"686b04b4de4e7912","entry":"filter_indices_by_diff","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/tasks/curriculum_task.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/tasks/curriculum_task.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"686b04b4de4e7912"}},{"code_sha256_prefix":"2456857cfcfb2b2a","entry":"filter_indices_by_length","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/tasks/curriculum_task.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/tasks/curriculum_task.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"2456857cfcfb2b2a"}},{"code_sha256_prefix":"afe05e8787d26825","entry":"fscore","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/utils.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"afe05e8787d26825"}},{"code_sha256_prefix":"aeee078a872841dc","entry":"label_smoothed_nll_loss","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/criterions/qe_loss.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/criterions/qe_loss.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"aeee078a872841dc"}},{"code_sha256_prefix":"463dec4b9e606031","entry":"precision","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/utils.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"463dec4b9e606031"}},{"code_sha256_prefix":"758494073a00f760","entry":"precision","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/criterions/qe_metric.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/criterions/qe_metric.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"758494073a00f760"}},{"code_sha256_prefix":"1ac84079458cd9bf","entry":"precision_recall_fscore_support","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/criterions/qe_metric.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/criterions/qe_metric.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"1ac84079458cd9bf"}},{"code_sha256_prefix":"a10d395718b0825a","entry":"recall","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/utils.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/utils.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"a10d395718b0825a"}},{"code_sha256_prefix":"4cb708962c259bca","entry":"scores_for_class","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/criterions/qe_metric.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/criterions/qe_metric.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"4cb708962c259bca"}},{"code_sha256_prefix":"597e001a2b1cb7b2","entry":"select_positions","repo":"NJUNLP/njuqe","repo_kind":"found_in_text","path":"njuqe/models/xlmqe.py","file_url":"https://github.com/NJUNLP/njuqe/blob/HEAD/njuqe/models/xlmqe.py","link_basis":"harvester_set","language":"python","status":"unverified","verification_level":0,"contract_check":null,"metamorphic_tier":null,"behaviour_fingerprint":false,"licence":"MIT","inline_ok":true,"mcp_get_code":{"code_sha256":"597e001a2b1cb7b2"}}]},"arxiv_metadata":null,"syntology_extracted_results":null}