Ying Li
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Ying Li has a total of 17 co-authors. Here are the top co-authors by number of shared publications:
- Yuxin Huang
School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming, China
3 shared publications - Zhengtao Yu
Faculty of Information Engineering and Automation, Kunming University of Science and Technology, Kunming, China
3 shared publications - Hua Lai2 shared publications
Shengxiang Gao
2 shared publicationsCunli Mao
1 shared publication- Cunli Mao
Kunming University of Science and Technology
1 shared publication
Latest Publications
- Open Access
- 22 July 2026
- Published by Elsevier in SSRN Electronic Journal
Enhancing Southeast Asian Grammatical Error Correction with New Benchmarks and Optimized Layer-wise Fine-tuning
Large Language Models (LLMs) have achieved remarkable success in high-resource grammatical error correction (GEC). However, they still face significant challenges in Southeast Asian low-resource languages due to the scarcity of high-quality parallel corpora and the complexity of distinct grammatical structures. In this work, we propose a unified LLM framework for Southeast Asian multi-lingual GEC tasks efficiently by lexical-guided data augmentation and an optimized layer-wise fine-tuning approach. Specifically, we formulate linguistically-informed confusion sets to guide LLMs in generating synthetic corpora and ensure their quality by further leveraging dual constraints of affinity and diversity. Meanwhile, we propose an optimized layer-wise fine-tuning approach that accurately detects and activates language-specific neural layers to adapt a single LLM for multiple Southeast Asian GEC tasks, where a composite score of $L_2$ norm and cosine similarity is used to quantify language-specific layer importance. Experimental results on four benchmark datasets demonstrate that our optimized layer-wise fine-tuning approach can significantly improve the generalization capabilities of LLMs in Southeast Asian low-resource GEC tasks. Detailed experimental analyses demonstrate that our proposed benchmark dataset introduces a significant variety of authentic errors, improving the model's ability to understand the syntactic structures of Southeast Asian low-resource languages. In-depth visualization results demonstrate that our approach effectively identifies and prioritizes layers critical for linguistic processing, providing evidence for future fine-grained optimization of Southeast Asian GEC tasks.- 22 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessEnhancing Southeast Asian Grammatical Error Correction with New Benchmarks and Optimized Layer-wise Fine-tuning
Large Language Models (LLMs) have achieved remarkable success in high-resource grammatical error correction (GEC). However, they still face significant challenges in Southeast Asian low-resource languages due to the scarcity of high-quality parallel corpora and the complexity of distinct grammatical structures. In this work, we propose a unified LLM framework for Southeast Asian multi-lingual GEC tasks efficiently by lexical-guided data augmentation and an optimized layer-wise fine-tuning approach. Specifically, we formulate linguistically-informed confusion sets to guide LLMs in generating synthetic corpora and ensure their quality by further leveraging dual constraints of affinity and diversity. Meanwhile, we propose an optimized layer-wise fine-tuning approach that accurately detects and activates language-specific neural layers to adapt a single LLM for multiple Southeast Asian GEC tasks, where a composite score of $L_2$ norm and cosine similarity is used to quantify language-specific layer importance. Experimental results on four benchmark datasets demonstrate that our optimized layer-wise fine-tuning approach can significantly improve the generalization capabilities of LLMs in Southeast Asian low-resource GEC tasks. Detailed experimental analyses demonstrate that our proposed benchmark dataset introduces a significant variety of authentic errors, improving the model's ability to understand the syntactic structures of Southeast Asian low-resource languages. In-depth visualization results demonstrate that our approach effectively identifies and prioritizes layers critical for linguistic processing, providing evidence for future fine-grained optimization of Southeast Asian GEC tasks. - Open Access
- 13 July 2026
- Published by Elsevier in SSRN Electronic Journal
Enhancing Multilingual Dependency Parsing via Dynamic Linguistic Feature Extraction and Alignment
Multi-lingual dependency parsing aims to extract useful syntactic knowledge from multiple high-resource languages to improve the parsing performance of low-resource target languages. How to effectively extract linguistic commonality structures and reduce divergence feature interferences is still a key challenge. To alleviate these issues, we propose a dynamic linguistic feature extraction and alignment framework for multi-lingual dependency parsing. On the one hand, we construct multi-lingual parallel dependency trees through confidence-aware traditional parser and LLM collaboration, thus explicitly extracting cross-lingual aligned syntactic knowledge. On the other hand, we introduce a dynamic feature alignment network to selectively transfer useful high-resource syntactic features and suppress harmful ones.Experiments on four low-resource benchmarks demonstrate that our model consistently improves multi-lingual parsing baselines, achieving competitive or superior performance on all languages. Further analyses indicate that the proposed framework effectively captures transferable structural knowledge and reduces negative transfer, leading to better parsing performance in low-resource settings. Our codes and data are available at \url{https://github.com/**}
- 13 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessEnhancing Multilingual Dependency Parsing via Dynamic Linguistic Feature Extraction and Alignment
Multi-lingual dependency parsing aims to extract useful syntactic knowledge from multiple high-resource languages to improve the parsing performance of low-resource target languages. How to effectively extract linguistic commonality structures and reduce divergence feature interferences is still a key challenge. To alleviate these issues, we propose a dynamic linguistic feature extraction and alignment framework for multi-lingual dependency parsing. On the one hand, we construct multi-lingual parallel dependency trees through confidence-aware traditional parser and LLM collaboration, thus explicitly extracting cross-lingual aligned syntactic knowledge. On the other hand, we introduce a dynamic feature alignment network to selectively transfer useful high-resource syntactic features and suppress harmful ones.Experiments on four low-resource benchmarks demonstrate that our model consistently improves multi-lingual parsing baselines, achieving competitive or superior performance on all languages. Further analyses indicate that the proposed framework effectively captures transferable structural knowledge and reduces negative transfer, leading to better parsing performance in low-resource settings. Our codes and data are available at \url{https://github.com/**}
- 24 February 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 315
Improving cross-lingual dependency parsing via LLM-based transferring and self-optimizing synthetic data augmentation
- 24 February 2026
- Published by Elsevier in Expert Systems with Applications
- Vol. 315
Improving cross-lingual dependency parsing via LLM-based transferring and self-optimizing synthetic data augmentation
- Open Access
- 24 October 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Asian and Low-Resource Language Information Processing
- Vol. 24 (11) , 1-17
Entity-focused Chinese Spelling Correction: Dataset and Approach
Due to the strong representation capability of pre-trained language models, Chinese spelling correction models have significantly improved. However, pre-trained language models focus on contextualized information and treat all words equally, thus ignoring the entity information. In practical application, entity words are the most difficult part to handle in various artificial intelligence tasks, i.e., machine translation, optical character recognition, and automatic speech recognition. To address this issue, we first construct an entity-focused Chinese spelling correction (EFCSC) dataset , the first public spelling correction corpus to emphasize entity errors. Furthermore, we propose an entity knowledge injected language model (EKILM) designed for entity-focused spelling correction, which injects entity information into pre-trained language models, thus ensuring traditional spelling correction models pay more attention to entity words. Experiments on several benchmark datasets show that our proposed model outperforms all strong baseline models, leading to state-of-the-art results on all datasets. Extensive experiments and detailed analyses demonstrate that our proposed model enhances entity error correction ability without damaging the normal spelling correction performance. Our code and dataset will be released at https://github.com/DPloved/EFCSC to facilitate future research.
- 24 October 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Asian and Low-Resource Language Information Processing
- Vol. 24 (11) , 1-17
Open AccessEntity-focused Chinese Spelling Correction: Dataset and Approach
Due to the strong representation capability of pre-trained language models, Chinese spelling correction models have significantly improved. However, pre-trained language models focus on contextualized information and treat all words equally, thus ignoring the entity information. In practical application, entity words are the most difficult part to handle in various artificial intelligence tasks, i.e., machine translation, optical character recognition, and automatic speech recognition. To address this issue, we first construct an entity-focused Chinese spelling correction (EFCSC) dataset , the first public spelling correction corpus to emphasize entity errors. Furthermore, we propose an entity knowledge injected language model (EKILM) designed for entity-focused spelling correction, which injects entity information into pre-trained language models, thus ensuring traditional spelling correction models pay more attention to entity words. Experiments on several benchmark datasets show that our proposed model outperforms all strong baseline models, leading to state-of-the-art results on all datasets. Extensive experiments and detailed analyses demonstrate that our proposed model enhances entity error correction ability without damaging the normal spelling correction performance. Our code and dataset will be released at https://github.com/DPloved/EFCSC to facilitate future research.