检索结果-内蒙古大学图书馆

6～(th) international conference on Spoken language processing

作者： Sun-Ah Jun Miru Oh Dept. of Ling. UCLAU.S.A Dept. of East Asian Lgs. ＆ Cultures UCLA ＆ Dept. of EnglishYeojoo Institute of TechnologyKorea

＜正＞Foreign accents in second language （L2） production are caused by interference from the phonological system and phonetic realization of the speaker’s first language （L1）. including both segmental and prosodic features. this paper examines the intonation structure of Seoul Korean and its realization by American English speakers. Four English speakers of Korean. differing in fluency. and two Korean spenkers participaled in the experiment. Forty sentences were designed to test the realization of intonation patterns by varying the number of syllables within a word and a sentence, and by varying the conditions for the segment-tone interaction. Results show that, as with segmemal data. more advanced L2 speakers produce more native-like intonation patterns and prosodie structure than less advanced speakers. However, although advanced L2 speakers are better at grouping words into phrases, they are not better at producing surface tonal realizations of an accentual phrase than less advanced speakers. this suggests that phonological properties of intonation are acquired earlier than phonetic properties of intonation.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Wiki-based Collaborative Learning For Colleges Specialty English Writing Education

Wiki-based Collaborative Learning For Colleges Specialty Eng...

引用

the 6th international conference on Computer Science & Education(第六届国际计算机新科技与教育学术会议 ICCSE 2011)

作者： Deti Xie Guoxiang Tang Xuan Wang College of Resources and Environment Southwest University Beibei ChongqingP.R.China College of Computer and Information science Southwest University Beibei ChongqingP.R.China

Specialty English writing education, which aims to help students to train and develop students literacies and skills required for successful communication in English academic and professional contexts, is very important in china university students’ academic studies. As witnessed by the popularity of sociocultural and constructivist approaches in language teaching, collaboration can play a vital role in helping learners to build knowledge and acquire skills. Over past decades although a lot of advanced technologies and learning platforms have been introduced to this field, most of them however just focused on interactive, personalized instruction effectiveness and little attention has been given to examine how collaborative writing can be offered, coordinated and assessed. As one of the most popular web2.0 technologies, wiki is believed to have large potential for fostering collaborative group work and creating learning resources. In this study, a Wiki-based collaborative writing environment for one college specialty English course entitled “E-Commerce Specialty English” was present and practiced. three important learning strategies including collaboration and knowledge sharing, peer assessment and learning process monitoring were adopted to assist the development of a collaborative writing community. Evaluation results showed that Wiki-based collaborative writing can effectively contribute to students writing participants, facilitate group work and make students pay more attention to readers.

关键词： Wiki Collaborative learning Speciality English writing Evaluation

来源：评论

学校读者我要写书评

暂无评论

Code Generation Based on Quality-Aware Reordering

Code Generation Based on Quality-Aware Reordering

引用

6th international conference on Intelligent Computing and Signal processing (ICSP)

作者： Yuan Song Luo Yu College of Computer Science and Technology Wuhan University of Science and Technology Wuhan China Hubei Provincial Key Laboratory of Intelligent Information Processing and Real-time Industrial Systems Wuhan China

ISBN: (数字)9798350376548

ISBN: (纸本)9798350376555

Translation efforts from natural language to code generation have achieved great success with large pre-trained language models. However, selecting the correct code output for the problem posed by the generated set of candidate codes is still challenging. In this paper, we address this problem by proposing an outcome-based quality-aware reordering method for code selection. this method scores each code by whether the set of candidate codes passes the test cases or not as a reference to select the top k best code outputs. In addition, this method depends on the quality and number of test cases. However, it is impractical to manually create a large number of test cases for each problem, so this paper adopts an automated approach that utilizes the current model to generate test cases for each problem. the experimental results show that the approach proposed in this paper obtains a significant improvement in the pass@k benchmark for each model, and the correctness and quality of test cases generated automatically using the model are discussed experimentally. Summarizing the experimental results and discussions, the result-based quality-aware reordering method proposed in this paper achieves a significant improvement in solving the problem of selecting the correct output from natural language to code generation, effectively improves the accuracy of evaluation by automating the generation of test cases, and provides an effective solution to improving the correctness of code generation.

关键词： Codes Accuracy Computational modeling Estimation Signal processing Benchmark testing

来源：评论

学校读者我要写书评

暂无评论

Optimizing Medical Image Report Generation with Varied Attention Mechanisms

Optimizing Medical Image Report Generation with Varied Atten...

引用

international conference on Contemporary Computing and Informatics (IC3I)

作者： Pochamreddy Mukesh Reddy Vivek Kumar Verma Mudunuri Venkata Chaitanya Varma Computer and Communication Engineering Manipal University Jaipur Jaipur India Department of Information Technology Manipal University Jaipur Jaipur India Data Science and Engineering Manipal University Jaipur Jaipur India

Image captioning plays a crucial role in interpreting visual information in the medical field. this research paper represents an innovative approach to produce descriptive and insightful reports for its respective medical images. the combination of natural language processing and computer vision techniques are used, an encoder-decoder architecture is employed which aids in leveraging by using a qualified InceptionV3 model to extract image features. the captions which are generated were based on a deep learning model with a Gated Recurrent Unit-based decoder and an attention mechanism, guaranteeing the connection between the generated text to the image content. the research focuses mainly on the dataset which consists of medical images with its associated reports by utilizing data extraction and preprocessing methods to enhance the accuracy and contextual understanding of the generated report. the experiments and results section demonstrate the implementation details and efficacy of the proposed model in generating meaningful report for the medical image, thereby facilitating enhanced interpretation and analysis of medical images. this research contributes to the advancement of image captioning techniques in the medical domain and holds significance for clinical decision support systems and medical image analysis.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Research on Text Classification and Topic Extraction Model of Medical English Corpus Based on Natural language processing

Research on Text Classification and Topic Extraction Model o...

引用

IEEE international conference on Power, Intelligent Computing and Systems (ICPICS)

作者： Xinli Zhang Lingyue Xie Guangdong Medical University Dongguan China South China University of Technology Guangzhou China

ISBN: (数字)9798350374315

ISBN: (纸本)9798350374322

In the medical field, a large amount of professional literature stores a wealth of knowledge, and effectively organizing and classifying this information is of great significance for medical research and practice. this paper proposes a deep learning-based text classification and topic extraction model for processing medical English corpora. We first obtained a corpus of 1000 medical texts, and then used natural language processing techniques to preprocess these texts. We propose a convolutional neural network (CNN) based text classification model that is capable of automatically learning and extracting features from text. Meanwhile, we utilize Latent Dirichlet Allocation (LDA) model for topic extraction. On the test set, our model outperforms other common text classification models such as Naive Bayes, Support Vector Machines, and Logistic Regression in both accuracy and F1-score. the research in this paper provides an effective method for the automatic classification and topic extraction of medical English texts, which has wide practicability and application value.

关键词： Support vector machines Computational modeling Text categorization Neural networks Feature extraction Natural language processing Data models Convolutional neural networks Resource management Text processing

来源：评论

学校读者我要写书评

暂无评论

Simulating Learning from language and Examples 24th

Simulating Learning from Language and Examples

引用

24th international conference on Artificial Intelligence in Education (AIED)

作者： Weitekamp, Daniel Rachatasumrit, Napol Wei, Rachael Harpstead, Erik Koedinger, Kenneth Carnegie Mellon Univ Pittsburgh PA 15289 USA Univ Illinois Champaign IL 61820 USA

ISBN: (纸本)9783031363351;9783031363368

Simulations of human learning can be used as computational models for evaluating theories of learning. they can also be taught interactively to author intelligent tutoring systems. Prior simulated learner systems have learned inductively from worked examples and correctness feedback. this work introduces a mechanism where simulated learners can also learn from natural language. Using a neural grammar parser with additional symbolic processing steps, we simulate the production of loose interpretations of verbal instructions. these interpretations can be combined with worked examples to resolve the ambiguities of either form of instruction alone. We find that our system has practical benefits over an alternative method using github Copilot and slightly better accuracy.

关键词： Learning Simulations Computational Models of Learning Natural language processing Code Synthesis Authoring Tools

来源：评论

学校读者我要写书评

暂无评论

High performance SOAP processing driven by data mapping template

引用

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 2006年 4025卷 152-168页

作者： Jun, Wei Lei, Hua Chunlei, Niu Haoran, Zheng Technology Center of Software Engineering Institute of Software Chinese Academy of Sciences Beijing100080 China

ISBN: (纸本)3540351264

web Services, with loosely-coupled, high-interoperable and platformindependent characteristics, is gaining popularity in distributed computing. However, web services suffers performance penalty because its protocol stack is based on XML. SOAP is used to specify wire message format in web services, and SOAP processing largely affects the performance of web services. In this paper, we firstly analyze the performance of web services on Java platform, and identify that data model mapping between XML data and Java data is the main impact factor on performance. therefore, we propose a new scheme of data model mapping - "Dynamic Early Binding" which enables to improve SOAP processing by avoiding Java reflection operations and proactively generating processing codes. this dynamic early binding is realized by Data Mapping Template (DMT), which is specified by extended context free grammar and implemented by pushdown automaton with output. We introduce the technique into our developed SOAP engine – SOAPExpress. the effectiveness is illustrated by yielding over 100% speedups compared to Apache Axis 1.2 in our benchmark. © IFIP international Federation for information processing 2006.

关键词： web services

来源：评论

学校读者我要写书评

暂无评论

Recuperative Reference Search Aided Approach Towards Voice Cloning

Recuperative Reference Search Aided Approach Towards Voice C...

引用

international conference on Computational Intelligence and Networks (CINE)

作者： Premanand Ghadekar Ojas Joshi Jignesh Barhate Basun Kundu Arman Nandeshwar Manthan Gujar Department of Information Technology Vishwakarma Institute of Technology Pune

ISBN: (数字)9798331516796

ISBN: (纸本)9798331516802

this proposed system presents a novel approach to enhancing synthesized speech by infusing emotional nuances. By fine-tuning existing models, emotional inflections and intonations are introduced, improving perceptual evaluation of speech quality (MFCC) and mean opinion scores (MOS). the paper presents the solution of cloning life-like voice of any individual to an astounding similarity to that of original which is achieved in vocal data as small as 10 minutes in duration. the paper sees the contributions made such as use of deep invariant optimal quantization to effectively condense neural network models without compromising their performance, surpassing results in the field of RPA, RCA and OA by factors of $\mathbf{6. 2 1 \%}$, 6.9% and 2.12% respectfully with the added. the collaborative effort across machine learning, natural language processing (NLP), and audio engineering promises high-quality audio indistinguishable from human speech, with applications spanning assistive technologies to customer service solutions. this research contributes to advancing voice cloning technology, ensuring synthesized speech integrates emotional depth and authenticity effectively.

关键词： Quantization (signal) Customer services Neural networks Cloning Collaboration Machine learning Speech enhancement Natural language processing Mel frequency cepstral coefficient Computational intelligence

来源：评论

学校读者我要写书评

暂无评论

Word Sense Disambiguation for Assamese

Word Sense Disambiguation for Assamese

引用

IEEE international Advance Computing conference

作者： Jumi Sarmah Shikhar Kr. Sarma Dept of Information Technology Gauhati University

ISBN: (纸本)9781467382878

Word Sense Disambiguation (WSD) is the process of identifying the proper sense of an ambiguous word depending on the particular context. It is to find the accurate sense s_i among the set of senses {s_1, s_2, ..., s_n}. this task was motivated by its interpretation in various Natural language processing (NLP) applications like IR, MT, QA, TC, SP etc. In this paper, machine learning technique - Naive Bayes Classifier was used for automatic disambiguation task. Training data was prepared with sense annotated features. For preparing sense annotated data we took help of the sense inventory. Currently, about 160 ambiguous words are present in the sense inventory derived from 18K and 25K words from Assamese Corpus and WordNet. the system is implemented in two phases. In the first phase, a total of 2.7K sense annotated training data and 800 test data were taken and a result of 71% accuracy was found. Analyzing the result depicts that accuracy improves as the training data size gradually increases and by the learned model generated in the previous iteration. In second phase we manually validate the outcomes of first-phase and we add those clean sense tagged data to previous training data set. than we train our system with our incresing training data (3.5K) which enhance the result accuracy. An iterative learning is adopted by the system and more accuracy of 7% is achieved. this paper aims to implement Assamese WSD system by NB classifier using lexical features and enhancement of the baseline method turns out in improving the classifier accuracy to 78%.

关键词： WSD Naive Bayes Classifier (NB) WordNet Corpus

来源：评论

学校读者我要写书评

暂无评论

JAPANESE-TO-CHINESE SPOKEN language TRANSLATION BASED ON thE SIMPLE EXPRESSION

JAPANESE-TO-CHINESE SPOKEN LANGUAGE TRANSLATION BASED ON THE...

引用

6～(th) international conference on Spoken language processing

作者： Yumi WAKITA Kenji MATSUI Advanced Technology Research Laboratories Matsushita Electric Industrial Co. Ltd. Kyoto vJapan Advanced Technology Research Laboratories Matsushita Electric Industrial Co. Ltd. Kyoto vJapan

this paper describes a Japanese-to-Chinese spoken language translation (SLT) method based on simple expression and presents the experimental results. the method is aimed at developing a compact speech translation system, which is robust for spontaneous spoken language phenomena, including the recognition errors and different expression from various speakers, the idea of translation method based on simple expression is that the mechanism interprets speech-act rather than the direct translation of the speaker's words. the method is realized by mapping the simple expression instead of deep parsing. In this method, only keywords in speech recognition results are extracted, and the corresponding target sentences are extracted from the database by selecting similar example sentences with the keywords. All training sentences in the bilingual corpus are rewritten into the simple expression and grouped by rules. the rules are specially designed to develop the compact size example database for translation. Comparing with other example-based approaches to SLT, the method based on simple expression is easy to realize, and especially it is practical to develop the SLT systems limited in specific domains.

关键词： JAPANESE-TO-CHINESE SPOKEN language TRANSLATION BASED ON thE SIMPLE EXPRESSION

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：