检索结果-内蒙古大学图书馆

2023 IEEE International Conference on Control, Electronics and Computer Technology, ICCECT 2023

作者： Luan, Xin Zhang, Jinwei Xu, Miaomiao Silamu, Wushour Li, Yanbing Xinjiang University Xinjiang Laboratory of Multi-Language Information Technology Xinjiang University Xinjiang Multilingual Information Technology Research Center Xinjiang University College of Information Science and Engineering Xinjiang China

ISBN: (纸本)9798350399028

Synthetic datasets alleviate the shortage of label data in the real world to some extent. But, synthetic datasets still have problems with the complexity of picture backgrounds and text diversity. It is well known that collecting large amounts of real data is a job that requires a lot of human resources and material resources. Therefore, we propose a small batch data augmentation strategy, hoping to improve significant performance by collecting and labeling small batches of real data. We have verified our ideas on a strong baseline. The result shows that the accuracy of the model can be significantly improved by replacing the synthetic dataset with the real dataset, which proved that real datasets could train the model better than synthetic datasets. Then, we use different enhancement strategies to expand the data of small batches of real data sets and explore the performance improvement of the model under the condition of low-resource real data. Finally, we mixed the augmented small batch of real datasets and synthetic datasets to make the model learn the image features of real scenes more elegantly. The results show that the proposed strategy can well fill the gap between synthetic and real datasets and improve the model performance. © 2023 IEEE.

关键词： Character recognition

来源：评论

学校读者我要写书评

暂无评论

WIP: Investigating the Use of AI Chatbots by Undergraduate Computer Science Students

WIP: Investigating the Use of AI Chatbots by Undergraduate C...

引用

Frontiers in Education (FIE) Conference

作者： Peter Ilic Center for Language Research School of Comp. Sc & Engineering University of Aizu Aizu-Wakamatsu Japan

ISBN: (数字)9798350351507

ISBN: (纸本)9798350363067

This work in progress paper investigates the use of AI chatbots by undergraduate computer science students learning English as a second language in Japan. The study employs a mixed-methods approach, combining qualitative surveys and interviews with quantitative clustering analysis. The objectives are to identify the types of AI chatbots used, determine their usage patterns, and explore the benefits and challenges associated with their use in language learning. The qualitative data collection (surveys, n=96) has been completed, while interviews and quantitative analysis are ongoing. The study aims to identify distinct clusters of AI chatbot users and their characteristics, highlight challenges, and contribute to the growing knowledge on AI application in STEM students studying a second language. Future research should explore long-term effects, optimal balance between AI-assisted and human-led instruction, and guidelines for integrating AI chatbots in language learning curricula.

关键词： Surveys Computer science Statistical analysis Education Oral communication Chatbots Object recognition Artificial intelligence Interviews Guidelines

来源：评论

学校读者我要写书评

暂无评论

Enhancing Automatic Speech Assessment Leveraging Heterogeneous Features and Soft Labels For Ordinal Classification

Enhancing Automatic Speech Assessment Leveraging Heterogeneo...

引用

IEEE Spoken language Technology Workshop

作者： Wen-Hsuan Peng Sally Chen Berlin Chen Department of Computer Science and Information Engineering National Taiwan Normal University The Language Training & Testing Center

ISBN: (数字)9798350392258

ISBN: (纸本)9798350392265

The general goal of automated speech assessment (ASA) is to provide a consistent and objective evaluation on the spoken language proficiency of an L2 learner or test-taker. In contrast to most previous work that treats ASA as a nominal multi-classification task and thus neglects the sequential nature of proficiency grades, this paper explores the notion of soft labels for use in ASA. In particular, we strive to enhance ASA performance by examining two critical issues: (1) the impact of applying soft labels instead of hard labels in the optimization of ordinal classification for ASA, and (2) the effects of combining self-supervised learning (SSL) with handcrafted indicator features via a novel modeling paradigm. Our results demonstrate that the proposed model can considerably enhance performance compared to existing strong baselines. The improvement is evident not only in the test dataset of seen prompts but also in those of unseen prompts, suggesting the robust generalization and adaptability of our method.

关键词： Adaptation models Conferences Neural networks Self-supervised learning Speech enhancement Optimization

来源：评论

学校读者我要写书评

暂无评论

Tourism named entity recognition method based on knowledge enhancement 5

Tourism named entity recognition method based on knowledge e...

引用

5th International Conference on Artificial Intelligence and Computer Science, AICS 2023

作者： Leng, Tianlan Altenbek, Gulila Ma, Yajing Haisa, Gulzada College of Information Science and Engineering Xinjiang University Xinjiang Urumqi830017 China The Base of Kazakh and Kirghiz Language of National Language Resource Monitoring and Research Center on Minority Languages Xinjiang University Xinjiang Urumqi830017 China Xinjiang Multilingual Information Technology Laboratory Xinjiang Urumqi830017 China

ISBN: (纸本)9781510668621

Aiming at the problems of multiple meanings, long lengths, and close connections with context information in the recognition of named entities in the field of Chinese tourism, as well as the complex composition of some entities, this paper proposes knowledge-enhanced tourism naming entity recognition method. Firstly, the knowledge-enhanced ERNIE pre-trained language model is utilized to obtain the semantic representation of tourism text. Secondly, the obtained word vectors perform feature learning and feature representation of local information on the input data through Convolutional Neural Networks (CNNs). Then, the Bi-directional Long Short-Term Memory (BiLSTM) is used to fully learn the forward and backward feature information of tourism text. At last, the tag decoding layer based on Conditional Random Fields (CRFs) is employed to address tag dependencies and produce the optimal sequence of tourism named entity tags. The effectiveness of the ERNIE-CNN-BiLSTM-CRF model is confirmed by experimental results, which involve a comparison with other models using the created tourism dataset. © 2023 SPIE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Legal long-text case retrieval method based on multi-scale feature fusion technology

Legal long-text case retrieval method based on multi-scale f...

引用

2024 International Conference on Computer Application and Information Security, ICCAIS 2024

作者： Wang, Yiwen Qi, Xiaoke Zhao, Xiaobing School of Information Engineering Minzu University of China Beijing100081 China National Language Resource Monitoring & Research Center of Minority Languages Beijing100081 China The Department of Science and Technology China University of Political Science and Law Beijing102249 China

ISBN: (纸本)9781510689312

Legal case retrieval aims to retrieve similar candidate cases from a candidate legal case database for a case pending judgment, serving as an important reference for the decision. This task is of great significance for maintaining the unity of the legal system. Traditional text retrieval methods mainly rely on keyword matching and fail to delve into the complex semantic level of legal texts. Although pre-trained models such as bidirectional encoder representations from transformers (BERT) perform well in semantic understanding, they have limitations on the length of input characters and cannot obtain the full-text semantics of legal documents. Therefore, this paper proposes a legal long-text case retrieval method that combines large language model summarization and multi-scale feature fusion technology. This method applies the BM25 algorithm to filter relevant sentences within the queried case segments and trains BERT models to filter sentences related to the queried case. Secondly, a large language model is used to generate summary paragraphs for candidate cases. Finally, by splicing and fusing paragraph-level and sentence-level features at two scales, the multi-scale semantic information of the query and candidate cases is enhanced. The structure-aware pre-trained language model for legal case retrieval (SAILER) is used to vectorize the candidate cases after multi-scale feature fusion, and cosine similarity is used to calculate the relevance scores between the query and candidate case texts. A multi-field fusion method is designed to calculate a comprehensive score and rank. This paper’s method was compared with three traditional algorithms-TF-IDF algorithm, LMIR algorithm, BM25 algorithm-and pre-trained model methods such as SAILER for case retrieval. The experimental results show that this paper’s method effectively improves the accuracy and efficiency of long-text case matching. © The Authors.

关键词： Online searching

来源：评论

学校读者我要写书评

暂无评论

Potential mechanism of Lactiplantibacillus plantarum FS5-5 in ameliorating lipid metabolism disorders induced by high-fat diet

引用

Food Science and Human Wellness 2025年

作者： Tong Li Junrui Wu Haibo Mu Yiming Liu Yanfeng Liu Yan Zhu Suning Xia Zhenmin Liu Rina Wu College of Food Science， Shenyang Agricultural University Engineering Research Center of Food Fermentation Technology(Liaoning) Key Laboratory of Microbial Fermentation Technology Innovation(Shenyang) State Key Laboratory of Dairy Biotechnology， Shanghai Engineering Research Center of Dairy Biotechnology， Dairy Research Institute， Bright Dairy & Food Co.， Ltd. College of Foreign Language Teaching， Shenyang Agricultural University

Probiotics exert numerous beneficial effects by regulating gut microbiota composition and are considered a promising nutritional strategy for ameliorating lipid metabolism disorders. This study evaluated the beneficial effects of Lactiplantibacillus plantarum FS5-5 (L. plantarum FS5-5)， a candidate probiotic with lipid-lowering ability， on diet-induced lipid metabolism disorders. The results showed that strain FS5-5 significantly decreased weight gain， fat coefficient， and lipid levels while alleviating liver oxidative stress damage in high-fat diet (HFD) mice. Serum metabolomics results showed that L. plantarum FS5-5 modulated lipid metabolism， particularly in linoleic acid metabolism， alpha-linolenic acid metabolism， and primary bile acids (BAs) biosynthesis pathways. Mechanism analysis suggested that L. plantarum FS5-5 promoted fatty acid β-oxidation (FAO) by regulating the expression of key genes in the lipid metabolism pathway， including Fxr， Fgf15， Shp， Pparγ， Pgc1α， etc. This effectively reduced several medium/long-chain acylcarnitines levels in serum and promoted fat breakdown. Further analysis showed that the abundance of gut-beneficial bacteria was increased， and intestinal permeability and barrier damage were alleviated after FS5-5 intervention. Spearman correlation analysis showed that several specific genera were notably associated with HFD-related indicators. These findings demonstrated that L. plantarum FS5-5 was a potential functional food for preventing lipid metabolism disorders.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Corpus Development for Pronoun Substitute and Address Term Study 8

Corpus Development for Pronoun Substitute and Address Term S...

引用

8th International Conference on Business and Industrial Research, ICBIR 2023

作者： Sornlertlamvanich, Virach Wittayapanyanon, Sunisa Musashino University Aaii Department of Data Science Tokyo Japan Thammasat University Faculty of Engineering Pathumthani Thailand Tokyo University of Foreign Studies World Language and Society Education Center Tokyo Japan

ISBN: (纸本)9798350399646

The development of a corpus of pronoun substitute and address term annotations is aimed at exploring the actual use of these language elements in dialogue sentences. To create the corpus, dialogue sentences were extracted from a collection of popular TV drama scripts and novels, representing both genres in order to provide a comprehensive overview of conversation settings. One key feature of the corpus is the inclusion of a detailed definition of kinship relations. This was done to gain a better understanding of the usage and meaning of kinship terms in dialogue. To ensure the accuracy of the corpus annotations, each sentence is annotated by two annotators, and the results are compared. This approach helps to minimize errors and personal biases, leading to a high-quality corpus that accurately reflects the usage of pronoun substitutes and address terms in natural language conversations. The comparison of annotations can also highlight any discrepancies, which can then be resolved through further discussion and collaboration between the annotators. This rigorous process helps to ensure that the corpus is a reliable and trustworthy resource. The resulting corpus will provide a wealth of information about these language elements and the way they are used in real-life dialogues, making it a useful resource for researchers in this field. © 2023 IEEE.

关键词： address term corpus development dialogue kinship term pronoun substitute

来源：评论

学校读者我要写书评

暂无评论

A Streamable Neural Audio Codec with Residual Scalar-Vector Quantization for Real-Time Communication

arXiv

引用

arXiv 2025年

作者： Jiang, Xiao-Hang Ai, Yang Zheng, Rui-Chen Ling, Zhen-Hua National Engineering Research Center of Speech and Language Information Processing University of Science and Technology of China Hefei230027 China

This paper proposes StreamCodec, a streamable neural audio codec designed for real-time communication. StreamCodec adopts a fully causal, symmetric encoder-decoder structure and operates in the modified discrete cosine transform (MDCT) domain, aiming for low-latency inference and real-time efficient generation. To improve codebook utilization efficiency and compensate for the audio quality loss caused by structural causality, StreamCodec introduces a novel residual scalar-vector quantizer (RSVQ). The RSVQ sequentially connects scalar quantizers and improved vector quantizers in a residual manner, constructing coarse audio contours and refining acoustic details, respectively. Experimental results confirm that the proposed StreamCodec achieves decoded audio quality comparable to advanced non-streamable neural audio codecs. Specifically, on the 16 kHz LibriTTS dataset, StreamCodec attains a ViSQOL score of 4.30 at 1.5 kbps. It has a fixed latency of only 20 ms and achieves a generation speed nearly 20 times real-time on a CPU, with a lightweight model size of just 7M parameters, making it highly suitable for real-time communication applications. Copyright © 2025, The Authors. All rights reserved.

关键词： Discrete cosine transforms

来源：评论

学校读者我要写书评

暂无评论

Computerized Diagnostic language Assessment of Oral Requesting-in-Interaction: Proof-of-Concept

Computerized Diagnostic Language Assessment of Oral Requesti...

引用

Information and Education Technology (ICIET), International Conference on

作者： Allan Nicholas Maxim Mozgovoy John Blake Center for Language Research University of Aizu Aizuwakamatsu Japan School of Computer Science and Engineering University of Aizu Aizuwakamatsu Japan

ISBN: (数字)9798331537845

ISBN: (纸本)9798331537852

In this paper, we provide a proof-of-concept for a computerized diagnostic language assessment (C-DLA) of spoken requesting among Japanese learners of English at a Japanese computer science university. The program focuses on the pragmatics aspect of spoken communication, in which the language choices we make are affected by socio-contextual factors such as relative social status and familiarity with the interlocutor. In so doing, the C-DLA aims to address a number of challenges. Requesting is an important, but challenging skill for many learners, and yet is also undertaught in the language classroom, due to time and resource constraints. Further, assessments typically evaluate the learner holistically, providing an overall score, without providing insight into specific aspects of the learner's performance. A C-DLA addresses these challenges by employing a three-stage process: i) assessment administration; ii) provision of immediate individualized feedback to the learner that promotes learning, and iii) further instruction provision, based on the assessment results. Computerization of the DLA widens access, allowing large numbers of learners to engage with the program simultaneously. Here, we provide a rationale for the C-DLA and an outline of the key challenges - namely, speech recognition of L2 English, identification of pragmatic inappropriateness in a learner's interactive speaking performance and automated feedback delivery. We further provide a proof-of-concept for the C-DLA, in which the program administers a number of interactive spoken requesting tasks to the learner, acts as an automated spoken dialogue interlocutor, and provides immediate, automated pragmatics-focused feedback when necessary.

关键词： Computer science Education Prototypes Speech recognition Time factors Iterative methods Pragmatics Testing Python Graphical user interfaces

来源：评论

学校读者我要写书评

暂无评论

APCodec+: A Spectrum-Coding-Based High-Fidelity and High-Compression-Rate Neural Audio Codec with Staged Training Paradigm

arXiv

引用

arXiv 2024年

作者： Du, Hui-Peng Ai, Yang Zheng, Rui-Chen Ling, Zhen-Hua National Engineering Research Center of Speech and Language Information Processing University of Science and Technology of China Hefei China

This paper proposes a novel neural audio codec, named APCodec+, which is an improved version of APCodec. The APCodec+ takes the audio amplitude and phase spectra as the coding object, and employs an adversarial training strategy. Innovatively, we propose a two-stage joint-individual training paradigm for APCodec+. In the joint training stage, the encoder, quantizer, decoder and discriminator are jointly trained with complete spectral loss, quantization loss, and adversarial loss. In the individual training stage, the encoder and quantizer fix their parameters and provide high-quality training data for the decoder and discriminator. The decoder and discriminator are individually trained from scratch without the quantization loss. The purpose of introducing individual training is to reduce the learning difficulty of the decoder, thereby further improving the fidelity of the decoded audio. Experimental results confirm that our proposed APCodec+ at low bitrates achieves comparable performance with baseline codecs at higher bitrates, thanks to the proposed staged training paradigm. Copyright © 2024, The Authors. All rights reserved.

关键词： Discriminators

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：