{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T23:53:11Z","timestamp":1774569191637,"version":"3.50.1"},"reference-count":16,"publisher":"Wiley","issue":"2","license":[{"start":{"date-parts":[[2026,2,16]],"date-time":"2026-02-16T00:00:00Z","timestamp":1771200000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"},{"start":{"date-parts":[[2026,2,16]],"date-time":"2026-02-16T00:00:00Z","timestamp":1771200000000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Internet Technology Letters"],"published-print":{"date-parts":[[2026,3]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                  <jats:p>Recent grammar error correction (GEC) systems have scaled rapidly in model size and architectural depth, creating a growing mismatch between algorithmic improvements and the latency and energy constraints of edge devices. The method reformulates English GEC as a task\u2010constrained latent editing problem, where grammatical corrections are represented as low\u2010rank perturbations in a compact linear subspace. A Tiny\u2010LM\u2010style weight re\u2010parameterization aligns the latent editing vectors with a minimal set of re\u2010parameterized weights, ensuring that English grammatical reasoning is concentrated in a hardware\u2010friendly linear manifold. To improve correction fidelity under tight computational budgets, a two\u2010stage progressive refinement strategy is employed: a fixed\u2010window lookahead performs coarse structural edits, followed by a sparse consistency filter that selectively verifies candidate token corrections under INT8\/INT4 quantization. The entire pipeline is static\u2010shape and operator\u2010regular, relying solely on linear, NPU\u2010native operations for predictable latency and bounded memory footprint. Experiments on public datasets show that the proposed model outperforms large Transformer baselines in F0.5 score on typical edge NPUs while reducing latency by 3\u20137\u00d7, demonstrating that accurate, low\u2010latency, on\u2010device English GEC is achievable using generic NPU operators without heavyweight language models.<\/jats:p>","DOI":"10.1002\/itl2.70240","type":"journal-article","created":{"date-parts":[[2026,2,17]],"date-time":"2026-02-17T00:00:26Z","timestamp":1771286426000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["A Compact Model for English Grammar Error Correction in the Low\u2010Latency Edge Deployment"],"prefix":"10.1002","volume":"9","author":[{"ORCID":"https:\/\/orcid.org\/0009-0009-7704-7265","authenticated-orcid":false,"given":"Shaoli","family":"Xiong","sequence":"first","affiliation":[{"name":"Dongguan City University  Guangdong China"},{"name":"UCSI University  Kuala Lumpur Malaysia"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"311","published-online":{"date-parts":[[2026,2,16]]},"reference":[{"key":"e_1_2_8_2_1","unstructured":"X.Sun T.Ge F.Wei andH.Wang Instantaneous Grammatical Error Correction With Shallow Aggressive Decoding. arXiv Preprint arXiv: 2106.04970(2021)."},{"key":"e_1_2_8_3_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2021.3077350"},{"key":"e_1_2_8_4_1","doi-asserted-by":"publisher","DOI":"10.1109\/IROS58592.2024.10801828"},{"key":"e_1_2_8_5_1","doi-asserted-by":"publisher","DOI":"10.1109\/TIP.2022.3205216"},{"key":"e_1_2_8_6_1","doi-asserted-by":"publisher","DOI":"10.1109\/JBHI.2025.3588551"},{"key":"e_1_2_8_7_1","doi-asserted-by":"publisher","DOI":"10.1109\/TCE.2024.3513331"},{"key":"e_1_2_8_8_1","article-title":"Compact re\u2010Parameterized Architectures for Efficient Language and Vision Modeling","volume":"103","author":"Zhang W.","year":"2025","journal-title":"Information Fusion"},{"key":"e_1_2_8_9_1","doi-asserted-by":"crossref","unstructured":"K.Omelianchuk V.Atrasevych A.Chernodub andO.Skurzhanskyi GECToR\u2013Grammatical Error Correction: Tag Not Rewrite. arXiv Preprint arXiv:2005.12592(2020).","DOI":"10.18653\/v1\/2020.bea-1.16"},{"key":"e_1_2_8_10_1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00286"},{"key":"e_1_2_8_11_1","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/W19-4406"},{"key":"e_1_2_8_12_1","doi-asserted-by":"publisher","DOI":"10.3115\/v1\/W14-1701"},{"key":"e_1_2_8_13_1","first-page":"180","volume-title":"Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies","author":"Yannakoudakis H.","year":"2011"},{"key":"e_1_2_8_14_1","first-page":"110","volume-title":"Proceedings of the 31st International Conference on Computational Linguistics","author":"Deng J.","year":"2025"},{"key":"e_1_2_8_15_1","first-page":"10323","volume-title":"SparseGPT: Massive Language Models Can Be Accurately Pruned in One\u2010Shot","author":"Frantar E.","year":"2023"},{"key":"e_1_2_8_16_1","doi-asserted-by":"crossref","unstructured":"N.Balepur M.Shu A.Hoyle et al. A Smart Mnemonic Sounds Like \u201cGlue Tonic\u201d: Mixing LLMs With Student Feedback to Make Mnemonic Learning Stick. arXiv Preprint arXiv:2406.15352(2024).","DOI":"10.18653\/v1\/2024.emnlp-main.786"},{"key":"e_1_2_8_17_1","first-page":"15801","article-title":"Stress\u2010Testing Long\u2010Context Language Models With Lifelong ICL and Task Haystack","volume":"37","author":"Xu X.","year":"2024","journal-title":"Advances in Neural Information Processing Systems"}],"container-title":["Internet Technology Letters"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/itl2.70240","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1002\/itl2.70240","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1002\/itl2.70240","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,3,26]],"date-time":"2026-03-26T23:06:54Z","timestamp":1774566414000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1002\/itl2.70240"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,16]]},"references-count":16,"journal-issue":{"issue":"2","published-print":{"date-parts":[[2026,3]]}},"alternative-id":["10.1002\/itl2.70240"],"URL":"https:\/\/doi.org\/10.1002\/itl2.70240","archive":["Portico"],"relation":{},"ISSN":["2476-1508","2476-1508"],"issn-type":[{"value":"2476-1508","type":"print"},{"value":"2476-1508","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,16]]},"assertion":[{"value":"2025-11-17","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-02-05","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-02-16","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70240"}}