检索结果-内蒙古大学图书馆

Chinese Grammatical Error Correction via Large Language Model Guided Optimization Training 23

23rd Chinese National Conference on Computational Linguistics, CCL 2024

作者： Liu, Xiao Li, Ying Yu, Zhengtao Faculty of Information Engineering and Automation Kunming University of Science and Technology Kunming China Yunnan Key Laboratory of Artificial Intelligence Kunming China

ISBN: (纸本)9780000000002

Pre-trained language model-based methods for Chinese Grammatical Error Correction (CGEC) are categorized into Seq2Seq and Seq2Edit types. However, both Seq2Seq and Seq2Edit models depend on high-quality training data significantly. Considering the strong generation and inference ability of large language models (LLMs), we propose a large language model-guided optimization training method to exploit LLMs to extract error knowledge to optimize the traditional CGEC model training process. On the one hand, we use error types and confusion sets as extra knowledge to guide LLMs to generate diverse pseudo data, thus extending the error distribution of our training data. On the other hand, LLMs are utilized to infer the predicted results from our CGEC models and obtain the re-training data, thus iteratively optimizing our pre-trained CGEC models. Experiments on two benchmark datasets show that our LLMs-guided optimization method with small-scale training data can achieve comparable results with baseline models with large-scale training data. Detailed comparison experiments demonstrate that both the early deviser pseudo data and the later re-training data are extremely useful for traditional CGEC model optimization training, and can benefit from each other. We will release our code and prompts at https://***/SakuraAcedia/llm-cgec-got to facilitate future work. © 2024 China National Conference on Computational Linguistics.

关键词： Digital elevation model

来源：评论

学校读者我要写书评

暂无评论

Multi-features Enhanced Multi-task Learning for Vietnamese Treebank Conversion 23rd

Multi-features Enhanced Multi-task Learning for Vietnamese ...

引用

23rd China National Conference on Computational Linguistics, CCL 2024

作者： Zhang, Zhenguo Liu, Jianjian Li, Ying Yunnan Key Laboratory of Artificial Intelligence Kunming Kunming650500 China Faculty of Information Engineering and Automation Kunming University of Science and Technology Kunming650500 China

ISBN: (纸本)9789819783663

Pre-trained language representation-based dependency parsing models have achieved obvious improvements in rich-resource languages. However, these model performances depend on the quality and scale of training data significantly. Compared with Chinese and English, the scale of Vietnamese Dependency treebank is scarcity. Considering human annotation is labor-intensive and time-consuming, we propose a multi-features enhanced multi-task learning framework to convert all heterogeneous Vietnamese Treebanks to a unified one. On the one hand, we exploit Tree BiLSTM and pattern embedding to extract global and local dependency tree features from the source Treebank. On the other hand, we propose to integrate these features into a multi-task learning framework to use the source dependency parsing to assist the conversion processing. Experiments on the benchmark datasets show that our proposed model can effectively convert heterogeneous treebanks, thus further improving the Vietnamese dependency parsing accuracy by about 7.12 points in LAS. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Multi-task learning

来源：评论

学校读者我要写书评

暂无评论

Dynamic Syntactic Feature Filtering and Injecting Networks for Cross-lingual Dependency Parsing 39

Dynamic Syntactic Feature Filtering and Injecting Networks f...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Liu, Jianjian Yu, Zhengtao Li, Ying Huang, Yuxin Gao, Shengxiang Yunnan Provincial Key Laboratory of Artificial Intelligence Faculty of Information Engineering and Automation Kunming University of Science and Technology Kunming650000 China

ISBN: (纸本)157735897X

Pre-trained language models enhanced parsers have achieved outstanding performance in rich-resource languages. Cross-lingual dependency parsing aims to learn useful knowledge from high-resource languages to alleviate data scarcity in low-resource languages. However, effectively reducing the syntactic structure distributional bias and excavating the commonalities among languages is the key challenge for cross-lingual dependency parsing. To address this issue, we propose novel dynamic syntactic feature filtering and injecting networks based on the typical shared-private model that employs one shared and two private encoders to separate source and target language features. Concretely, a Language-Specific Filtering Network (LSFN) on private encoders emphasizes helpful information and ignores the irrelevant or harmful parts of it from the source language. Meanwhile, a Language-Invariant Injecting Network (LIIN) on the shared encoder integrates the advantages of BiLSTM and improved Transformer encoders to transcend language boundaries, thus amplifying syntactic commonalities across languages. Experiments on seven benchmark datasets show that our model achieves an average absolute gain of 1.84 UAS and 3.43 LAS compared with the shared-private model. Comparative experiments validate that both LSFN and LIIN components are complementary in transferring beneficial knowledge from source to target languages. Detailed analyses highlight that our model can effectively capture linguistic commonalities and mitigate the effect of distributional bias, showcasing its robustness and efficacy. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Syntactics

来源：评论

学校读者我要写书评

暂无评论

Fault Diagnosis of Large-Scale Railway Maintenance Equipment Based on GA-RBF Neural Network 11th

Fault Diagnosis of Large-Scale Railway Maintenance Equipmen...

引用

11th EAI International Conference on Context-Aware Systems and Applications, ICCASA 2022

作者： Wang, Hairui Li, Yuanbo Zhang, Wenqi Duan, Yusu Zhu, Guifu Faculty of Information Engineering and Automation Kunming University of Science and Technology Kunming China Information Technology Construction Management Center Kunming University of Science and Technology Kunming China

ISBN: (纸本)9783031288159

At present, the large-scale railway maintenance equipment adopts a diesel engine as the main power plant. Therefore the diesel engine in the event of failure, will seriously affect the large-scale railway maintenance equipment of the normal work. Exploring advanced diesel engine condition monitoring and fault diagnosis technology and looking for practical and effective diesel engine fault diagnosis method, which has already become a research subject widely concerned by many experts at home and abroad. In this paper, genetic algorithm (GA) is used to optimize the parameters of radial basis function (RBF) neural network for diesel engine fault diagnosis, experimental results show the validity of this prediction method, and the accuracy of the proposed algorithm was verified by comparative. © 2023, ICST Institute for Computer sciences, Social Informatics and Telecommunications engineering.

关键词： Failure analysis

来源：评论

学校读者我要写书评

暂无评论

Multiframe Blind Deconvolution based on Bright and Dark Channel Prior

Multiframe Blind Deconvolution based on Bright and Dark Chan...

引用

2023 International Conference on Algorithms, High Performance Computing, and Artificial Intelligence, AHPCAI 2023

作者： Ai, Haotian Faculty of Information Engineering and Automation Kunming University of Science and Technology Kunming Yunnan650000 China

ISBN: (纸本)9781510671881

Blind image restoration is a challenging and valuable subject in the field of computer vision. In this paper, under the framework of Total Variation (TV) regularization term, the bright and dark channel regularization term is introduced into the multi-frame image blind restoration algorithm, aiming at overcoming the shortcomings of low contrast and obvious ringing effect of the restored image on the basis of retaining the advantages of TV regularization. The bright channel regularization term can estimate the blur degree of the degraded image more accurately, while the dark channel regularization term can better resist noise interference. The split Bregman iterative algorithm is used to solve the objective function, and the details of the restored image are compensated in the iterative process, and further. The experimental results show that the algorithm in this paper has a certain deblurring effect, and the peak signal-to-noise ratio of the restored image is improved by about 0.53dB~1.5dB, and the structural similarity is improved by about 0.12~0.47, with better visual effects. © 2023 SPIE. All rights reserved.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

BAMGAN: A KBQA method based GAN and bi-directional attention 6

BAMGAN: A KBQA method based GAN and bi-directional attention

引用

6th International Conference on Advanced Electronic Materials, Computers, and Software engineering, AEMCSE 2023

作者： Guo, Jiazhi Faculty of Information Engineering and Automation Kunming University of Science and Technology Kunming Yunnan650504 China

ISBN: (纸本)9781510668249

Knowledge-based question answering(KBQA) involves using knowledge base technology to generate answers to natural language processing(NLP) questions. KBQA is one of the most challenging tasks in the field of NLP. Answer selection plays a vital role in KBQA, as it requires selecting the correct answer from a pool of candidate answers. However, knowledge bases still struggle to identify the correct answer from multiple candidate answers. To address this issue, In this paper, we propose BAMGAN, a KBQA method based on generative adversarial networks. This method uses generative adversarial networks (GANs) to improve answer selection accuracy in deep neural networks. Experimental results show the effectiveness of the BAMGAN. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Stability and Thermal Property Optimization of Propylene Glycol-Based MWCNT Nanofluids

引用

Fluid Dynamics & Materials Processing 2023年第9期19卷 2399-2416页

作者： Xi Wang Shan Qing Zhumei Luo Yiqin Liu Zichang Shi Jiachen Li Faculty of Metallurgical and Energy Engineering Kunming University of Science and TechnologyKunming650093China Faculty of Media and Information Engineering Yunnan Open UniversityKunming650093China

Propylene glycol-based MWCNT(multi-walled carbon nanotubes)nanoﬂuids were prepared in the framework of a two-step method and by using a suitable PVP(polyvinyl pyrrolidone)*** BBD(Box-Behnken design)model was exploited to analyze 17 sets of experiments and examine the sensitivity of the absorbance to three parameters,namely the concentration of MWCNT,the SN ratio(mass ratio of carbon nanotubes to sur-factants)and the sonication *** results have revealed that,while the SN ratio and concentration of MWCNT have a strong effect on the absorbance,the inﬂuence of the sonication time is less *** sta-tistical method of analysis of variance(ANOVA)was further used to determine the F-and p-values of the *** experiments were run to validate this *** sample 2 was found to display the greatest absorbance,it was selected for stability monitoring as well as thermal conductivity and viscosity *** sample has been found to be stable;the viscosity decreased with increasing temperature;the addition of MWCNT nano-particles was more effective in improving the thermal conductivity of propylene glycol than other methods in the ***,the MWCNT nanoﬂuid based on propylene glycol exhibited higher thermal conductivity at low temperatures.

关键词： Propylene glycol thermal conductivity response surface methodology stability viscosity

来源：评论

学校读者我要写书评

暂无评论

Analysis of the Correlation between Serum 25 Hydroxyvitamin D and Blood Lipids in Middle-Aged and Elderly People in Anning District, kunming City 23

Analysis of the Correlation between Serum 25 Hydroxyvitamin ...

引用

15th International Conference on Bioinformatics and Biomedical technology, ICBBT 2023

作者： Sun, Zhiran Song, Chao Zhang, Hongjiang Lu, Chengxin Qin, Zhengyu Guo, Ling Xiong, Xin He, Jianfeng Faculty of Information Engineering and Automation Kunming University of Science and Technology China China

ISBN: (纸本)9798400700385

Dyslipidemia is a metabolic disorder that can increase the risk of cardiovascular diseases. This study aims to investigate the association between serum 25- hydroxyvitamin D(25(OH)D) levels and blood lipid abnormalities in middle-aged and elderly residents of Anning District, kunming City. It provides a theoretical basis for the prevention and treatment of cardiovascular diseases in the region. A total of 6,391 middle-aged and elderly individuals from the First People's Hospital of Anning City, Yunnan Province, were categorized based on serum 25(OH)D levels. Pearson correlation analysis, multiple linear regression, and other methods were used to examine the relationship between 25(OH)D and blood lipid abnormalities. The results showed that the Body Mass Index(BMI), Total Cholesterol(TC), and Triglyceride (TG) levels in the severely deficient group, deficient group, and insufficient group were significantly higher than those in normal group (P © 2023 ACM.

关键词： Diseases

来源：评论

学校读者我要写书评

暂无评论

An Improved Multi-Site Feature Detection Model for Knee Arthritis Based on YOLOv5s 3rd

An Improved Multi-Site Feature Detection Model for Knee Arth...

引用

3rd International Conference on Machine Learning, Cloud Computing and Intelligent Mining, MLCCIM 2024

作者： Liu, Yingli Huang, Yishan Cha, Yinqiu Gao, Ming Faculty of Information Engineering and Automation Kunming University of Science and Technology Yunnan Kunming China The First People’s Hospital of Yunnan Province The Affiliated Hospital of Kunming University of Science and Technology Yunnan Kunming China

ISBN: (纸本)9789819624676

Addressing the needs and challenges of multi-site feature detection for knee arthritis, this paper proposes an improved lightweight network model, MMG_YOLOv5s. This model aims to solve the problems of complex structures and massive parameter counts in traditional detection models, which hinder their deployment on edge devices. The M(1) and M(2) modules in the MMG_YOLOv5s model enhance the YOLOv5s network by adjusting its backbone and neck structures, decreasing the count of network parameters and compressing the model's size. This, in turn, improves computational efficiency and creates conditions for deployment on resource-constrained devices. The data for this experiment is sourced from the public dataset OAI and annotated under the guidance of professional doctors to construct our own dataset. The experimental outcomes indicate that the recognition accuracy of this model achieves 80.8%, representing an 11% enhancement in comparison to the original YOLOv5s model. More importantly, its parameter count is reduced by 71.27%. This improvement provides a solid foundation for the subsequent embedded deployment of the model into portable or mobile devices, promising to advance the intelligent and convenient process of knee arthritis detection. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Diseases

来源：评论

学校读者我要写书评

暂无评论

Design of high-transmittance anomalous refraction metasurface based on geometric phase 9

Design of high-transmittance anomalous refraction metasurfac...

引用

9th International Conference on Energy Materials and Electrical engineering, ICEMEE 2023

作者： Zhang, Li Sun, Jun Hu, Mingzhi Chang, Kairong Dai, Shuhao School of Information Engineering and Automation Kunming University of Science and Technology Yunnan Kunming650000 China

ISBN: (纸本)9781510672765

In this paper, a transmission metasurface based on a geometric phase is proposed. By designing and arranging the metasurface unit structure, the wavefront of the incident circularly polarized wave can be manipulated arbitrarily. The metasurface unit is composed of a typical "sandwich"structure, metal-dielectric-metal, and the top and bottom metal patterns are composed of C-shaped and rectangles. Using the geometric phase principle, the phase can be adjusted by rotating the metal structure at the operating frequency, and the cross-polarization transmission amplitude is as high as 90 %. By rotating the coding of the unit structure, a coding metasurface can be formed, resulting in abnormal deflection. The simulation results show that the designed coding metasurface can regulate the electromagnetic wave well at 0.63 THz, and the abnormal refraction angles are ± 50.1° and 30°, which well verifies the generalized Snell's law and has potential application value in the fields of lidar and so on. © COPYRIGHT SPIE. Downloading of the abstract is permitted for personal use only.

关键词： Refraction Terahertz radiation Design and modelling Wavefronts Structural design

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：