检索结果-内蒙古大学图书馆

Cost-Effective Knowledge Extraction Framework for Low-Resource Environments

IEEE ACCESS 2024年 12卷 60668-60681页

作者： Nam, Sangha Kim, Eun-Kyung NCSoft Language AI Lab Seongnam 13494 South Korea Daejeon Univ Dept Big Data & AI Daejeon 34520 South Korea

Extracting knowledge from texts is crucial for enriching everyday knowledge. Constructing a knowledge extraction environment requires comprehensive processes, such as data generation, data processing, and model and framework design. However, these processes require significant effort in low-resource environments where shared data are not published. Currently, there is no environment that can design an entire knowledge extraction framework and perform step-by-step experiments even with unlimited resources. Thus, this study proposes a method for building a cost-effective knowledge extraction environment. In particular, we present a low-cost, high-quality method for annotating a corpus for knowledge extraction, in which data sharing is unavailable. The dataset collected using this method improves the performance of knowledge-extraction system models. Specifically, the co-reference resolution and relation extraction performance were improved by 10% and 18.9%, respectively. Additionally, the entire knowledge extraction system was evaluated using sequential multitask learning, and the performance was improved by 5% as each trained model was introduced.

关键词： Crowdsourcing knowledge base knowledge extraction low-resource environment

来源：评论

学校读者我要写书评

暂无评论

A Virtual Knowledge Distillation via Conditional GAN

引用

IEEE ACCESS 2022年 10卷 34766-34778页

作者： Kim, Sihwan Hana Inst Technol Big Data & AI Lab Seoul 06133 South Korea

Knowledge distillation aims at transferring the knowledge from a pre-trained complex model, called teacher, to a relatively smaller and faster one, called student. Unlike previous works that transfer the teacher's softened distributions or feature spaces, in this paper, we propose a novel approach, called Virtual Knowledge Distillation (VKD), that transfers a softened distribution generated by a virtual knowledge generator conditioned on class label. A virtual knowledge generator is trained independently, but concurrently with a teacher, to mimic the teacher's softened distributions. Afterwards, when training a student, virtual knowledge generator can be exploited instead of the teacher's softened distributions or combined with the existing distillation methods in a straightforward manner. Moreover, with slight modifications, VKD can be utilized not only for the self-knowledge distillation method but also for the collaborative learning method. We compare our method with several representative distillation methods in various combinations of teacher and student architectures on the image classification tasks. Experimental results on various image classification tasks demonstrate that VKD show a competitive performance compared to the conventional distillation methods, and when combined with them, the performance is improved with a substantial margin.

关键词： Training Generators Knowledge engineering Bridges Generative adversarial networks Task analysis Collaborative work Image classification model compression knowledge distillation self-knowledge distillation collaborative learning conditional generative adversarial network

来源：评论

学校读者我要写书评

暂无评论

AUTOPREP: AN AUTOMATIC PREPROCESSING FRAMEWORK FOR IN-THE-WILD SPEECH data 49

AUTOPREP: AN AUTOMATIC PREPROCESSING FRAMEWORK FOR IN-THE-WI...

引用

49th IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Yu, Jianwei Chen, Hangting Bian, Yanyao Li, Xiang Luo, Yi Tian, Jinchuan Liu, Mengyang Jiang, Jiayi Wang, Shuai Tencent AI Lab Shenyang Peoples R China Shenzhen Res Inst Big data Shenzhen Peoples R China

ISBN: (纸本)9798350344868;9798350344851

Recently, the utilization of extensive open-sourced text data has significantly advanced the performance of text-based large language models (LLMs). However, the use of in-the-wild large-scale speech data in the speech technology community remains constrained. One reason for this limitation is that a considerable amount of the publicly available speech data is compromised by background noise, speech overlapping, lack of speech segmentation information, missing speaker labels, and incomplete transcriptions, which can largely hinder their usefulness. On the other hand, human annotation of speech data is both time-consuming and costly. To address this issue, we introduce an automatic in-the-wild speech data preprocessing framework (AutoPrep) in this paper, which is designed to enhance speech quality, generate speaker labels, and produce transcriptions automatically. The proposed AutoPrep framework comprises six components: speech enhancement, speech segmentation, speaker clustering, target speech extraction, quality filtering and automatic speech recognition. Experiments conducted on the open-sourced WenetSpeech and our self-collected AutoPrepWild corpora demonstrate that the proposed AutoPrep framework can generate preprocessed data with similar DNSMOS and PDNSMOS scores compared to several open-sourced TTS datasets. The corresponding TTS system can achieve up to 0.68 in-domain speaker similarity.1

关键词： Speeech data preprocessing speech enhancement speaker clustering

来源：评论

学校读者我要写书评

暂无评论

Distribution pattern analysis of passing and receiving locations during the 2022 FIFA World Cup using the measure of spatial entropy

引用

INTERNATIONAL JOURNAL OF PERFORMANCE ANALYSIS IN SPORT 2025年第1期25卷 129-143页

作者： Xiong, Xiuyuan Deng, Kuiying Wu, Jiale Cui, Yixiong Beijing Sport Univ China Sports Big Data Ctr Sch Sports Engn Informat Rd 48 Beijing 100084 Peoples R China Beijing Sport Univ China Sports Big Data Ctr Sch Sports Engn AI Sports Engn Lab Beijing Peoples R China

The study aimed to investigate the spatial distribution patterns of passing and receiving locations during the 2022 FIFA World Cup in different match stages, playing positions and team possession types. Eighteen types of spatial entropy were proposed to evaluate the heterogeneity of passing and receiving locations in 64 matches. T-tests of these entropies were made between teams at the group and knockout stages, and between distinct playing positions. The K-means clustering of each team's ball-possession per match was employed to divide all matches into two categories, and their differences in entropies were then analysed using one-way ANOVA. Results showed that the teams at the group and knockout stages had significant differences only in goalkeeper entropy (ES = 0.368). The spatial entropy of the goalkeeper exhibited significant differences in comparison to the other positions (ES: -6.71 similar to -4.69). Low ball-possession teams were significantly lower in 8 types of spatial entropies (ES: -1.02 similar to -0.41) compared to their high ball-possession counterparts. Furthermore, only two types of spatial entropies (ES: -0.58 similar to -0.54) differed significantly in the comparison between high and low ball-possession teams. These findings may provide coaches with guidance for implementing targeted defensive strategies and enhancing the team's passing performance.

关键词： Complexity science ball possession spatial distribution passing performance soccer football

来源：评论

学校读者我要写书评

暂无评论

Enhancing Unsupervised Anomaly Detection With Score-Guided Network

引用

IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS 2024年第10期35卷 14754-14769页

作者： Huang, Zongyuan Zhang, Baohua Hu, Guoqiang Li, Longyuan Xu, Yanyan Jin, Yaohui Shanghai Jiao Tong Univ MoE Key Lab Artificial Intelligence Shanghai 200240 Peoples R China Shanghai Jiao Tong Univ AI Inst Shanghai 200240 Peoples R China Ind & Commercial Bank China Big Data & AI Lab Shanghai 201206 Peoples R China

Anomaly detection plays a crucial role in various real-world applications, including healthcare and finance systems. Owing to the limited number of anomaly labels in these complex systems, unsupervised anomaly detection methods have attracted great attention in recent years. Two major challenges faced by the existing unsupervised methods are as follows: 1) distinguishing between normal and abnormal data when they are highly mixed together and 2) defining an effective metric to maximize the gap between normal and abnormal data in a hypothesis space, which is built by a representation learner. To that end, this work proposes a novel scoring network with a score-guided regularization to learn and enlarge the anomaly score disparities between normal and abnormal data, enhancing the capability of anomaly detection. With such score-guided strategy, the representation learner can gradually learn more informative representation during the model training stage, especially for the samples in the transition field. Moreover, the scoring network can be incorporated into most of the deep unsupervised representation learning (URL)-based anomaly detection models and enhances them as a plug-in component. We next integrate the scoring network into an autoencoder (AE) and four state-of-the-art models to demonstrate the effectiveness and transferability of the design. These score-guided models are collectively called SG-Models. Extensive experiments on both synthetic and real-world datasets confirm the state-of-the-art performance of SG-Models.

关键词： Anomaly detection autoencoder (AE) regularization scoring network unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

datastore Distillation for Nearest Neighbor Machine Translation

引用

IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING 2024年 32卷 807-817页

作者： Dai, Yuhan Zhang, Zhirui Du, Yichao Liu, Shengcai Liu, Lemao Xu, Tong Univ Sci & Technol China Anhui Prov Key Lab Big Data Anal & Applicat Hefei 230026 Anhui Peoples R China Tencent AI Lab Shenzhen 518071 Guangdong Peoples R China Agcy Sci Technol & Res Ctr Frontier AI Res Singapore 138632 Singapore

Nearest neighbor machine translation (i.e., kNN-MT) is a promising approach to enhance translation quality by equipping pre-trained neural machine translation (NMT) models with the nearest neighbor retrieval. Despite its great success, kNN-MT typically requires ample space to store its token-level datastore, causing kNN-MT to be less practical in edge devices or online scenarios. In this paper, inspired by the concept of knowledge distillation, we provide a new perspective to ease the storage overhead by datastore distillation, which is formalized as a constrained optimization problem. We further design a novel model-agnostic iterative nearest neighbor merging method for the datastore distillation problem to obtain an effective and efficient solution. Experiments on three benchmark datasets indicate that our approach not only reduces the volume of the datastore by up to 50% without significant performance degradation, but also outperforms other baselines by a large margin at the same compression rate. Another experiment conducted on WikiText-103 further demonstrates the effectiveness of our method in the language model task.

关键词： Nearest neighbor machine translation datastore distillation

来源：评论

学校读者我要写书评

暂无评论

M-Vec: Matryoshka Speaker Embeddings with Flexible Dimensions 16th

M-Vec: Matryoshka Speaker Embeddings with Flexible Dimensio...

引用

16th International Conference on Social Robotics, ICSR + InnoBiz 2024

作者： Wang, Shuai Zhu, Pengcheng Li, Haizhou Shenzhen Research Institute of Big Data Shenzhen China Shenzhen India Fuxi AI Lab NetEase Inc. Hangzhou China

ISBN: (纸本)9789819611508

Fixed-dimensional speaker embeddings have become the dominant approach in speaker modeling, typically spanning hundreds to thousands of dimensions. These dimensions are hyperparameters that are not specifically picked, nor are they hierarchically ordered in terms of importance. In large-scale speaker representation databases, reducing the dimensionality of embeddings can significantly lower storage and computational costs. However, directly training low-dimensional representations often yields suboptimal performance. In this paper, we introduce the Matryoshka speaker embedding, a method that allows dynamic extraction of sub-dimensions from the embedding while maintaining performance. Our approach is validated on the VoxCeleb dataset, demonstrating that it can achieve extremely low-dimensional embeddings, such as 8 dimensions, while preserving high speaker verification performance. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Speech recognition

来源：评论

学校读者我要写书评

暂无评论

Affective Visualization Design: Leveraging the Emotional Impact of data

引用

IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS 2024年第1期30卷 1-11页

作者： Lan, Xingyu Wu, Yanqiu Cao, Nan Fudan Univ Shanghai Peoples R China Res Grp Computat & AI Commun Inst Global Commun & Shanghai Peoples R China Tongji Univ Intelligent Big Data Visualizat Lab Beijing Peoples R China

In recent years, more and more researchers have reflected on the undervaluation of emotion in data visualization and highlighted the importance of considering human emotion in visualization design. Meanwhile, an increasing number of studies have been conducted to explore emotion-related factors. However, so far, this research area is still in its early stages and faces a set of challenges, such as the unclear definition of key concepts, the insufficient justification of why emotion is important in visualization design, and the lack of characterization of the design space of affective visualization design. To address these challenges, first, we conducted a literature review and identified three research lines that examined both emotion and data visualization. We clarified the differences between these research lines and kept 109 papers that studied or discussed how data visualization communicates and influences emotion. Then, we coded the 109 papers in terms of how they justified the legitimacy of considering emotion in visualization design (i.e., why emotion is important) and identified five argumentative perspectives. Based on these papers, we also identified 61 projects that practiced affective visualization design. We coded these design projects in three dimensions, including design fields (where), design tasks (what), and design methods (how), to explore the design space of affective visualization design.

关键词： Information Visualization Affective Design Visual Communication User Experience Storytelling

来源：评论

学校读者我要写书评

暂无评论

Training-Free OOD Object Detection Leveraging Pre-Trained Segmentation Model Competency 15

Training-Free OOD Object Detection Leveraging Pre-Trained Se...

引用

15th International Conference on Information and Communication Technology Convergence, ICTC 2024

作者： Yun, Kimin Song, Jeonghoon Bae, Yuseok ETRI Visual Intelligence Lab. Daejeon Korea Republic of Soonchunhyang University Department of Ai and Big Data Asan Korea Republic of

ISBN: (纸本)9798350364637

This paper addresses a method for detecting Out-of-Distribution (OOD) objects by leveraging pre-trained segmentation models. The research on OOD detection focuses on the ability to accurately identify and classify untrained classes as 'unknown.' Previous methods rely on strong augmentation to detect OOD objects. However, these augmentation-based methods assume specific characteristics of OOD objects and can suffer from overfitting due to the limited datasets. In this study, we propose a training-free OOD object detection approach that infers OOD objects from a pre-trained segmentation model without additional training. Specifically, our method combines two modules: rejection of regions confidently recognized by the model (inliers) and selection of masks that capture the characteristics of OOD objects (outliers). Furthermore, using a class-agnostic segmentation model, the probability of OOD objects is refined at the object level, enhancing performance. Our method shows competitive performance on datasets where OOD objects are encountered in autonomous driving contexts. Additionally, we show the utility of measuring the model's competency to recognize what it knows and doesn't know from the perspective of the pre-trained model. © 2024 IEEE.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

RefXVC: Cross-Lingual Voice Conversion With Enhanced Reference Leveraging

引用

IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING 2024年 32卷 4146-4156页

作者： Zhang, Mingyang Zhou, Yi Ren, Yi Zhang, Chen Yin, Xiang Li, Haizhou Chinese Univ Hong Kong Shenzhen Res Inst Big Data Sch Data Sci Shenzhen 518172 Peoples R China Natl Univ Singapore Dept Elect & Comp Engn Singapore 119077 Singapore ByteDance AI Lab Speech & Audio Team Shanghai 201103 Peoples R China

This paper proposes RefXVC, a method for cross-lingual voice conversion (XVC) that leverages reference information to improve conversion performance. Previous XVC works generally take an average speaker embedding to condition the speaker identity, which does not account for the changing timbre of speech that occurs with different pronunciations. To address this, our method uses both global and local speaker embeddings to capture the timbre changes during speech conversion. Additionally, we observed a connection between timbre and pronunciation in different languages and utilized this by incorporating a timbre encoder and a pronunciation matching network into our model. Furthermore, we found that the variation in tones is not adequately reflected in a sentence, and therefore, we used multiple references to better capture the range of a speaker's voice. The proposed method outperformed existing systems in terms of both speech quality and speaker similarity, highlighting the effectiveness of leveraging reference information in cross-lingual voice conversion.

关键词： Timbre Task analysis Feature extraction Speech recognition Training Speech enhancement data mining Cross-lingual voice conversion (XVC) multi-reference pitch normalization speaker embedding

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：