检索结果-内蒙古大学图书馆

38th conference on Neural Information processing Systems, NeurIPS 2024

作者： Liu, Chenxi Chen, Yongqiang Liu, Tongliang Gong, Mingming Cheng, James Han, Bo Zhang, Kun TMLR Group Hong Kong Baptist University Hong Kong Mohamed bin Zayed University of Artificial Intelligence United Arab Emirates Carnegie Mellon University United States The Chinese University of Hong Kong Hong Kong Sydney AI Centre The University of Sydney Australia The University of Melbourne Australia

Revealing the underlying causal mechanisms in the real world is the key to the development of science. Despite the progress in the past decades, traditional causal discovery approaches (CDs) mainly rely on high-quality measured variables, usually given by human experts, to find causal relations. The lack of well-defined high-level variables in many real-world applications has already been a longstanding roadblock to a broader application of CDs. To this end, this paper presents Causal representatiOn AssistanT (COAT) that introduces large language models (LLMs) to bridge the gap. LLMs are trained on massive observations of the world and have demonstrated great capability in extracting key information from unstructured data. Therefore, it is natural to employ LLMs to assist with proposing useful high-level factors and crafting their measurements. Meanwhile, COAT also adopts CDs to find causal relations among the identified variables as well as to provide feedback to LLMs to iteratively refine the proposed factors. We show that LLMs and CDs are mutually beneficial and the constructed feedback provably also helps with the factor proposal. We construct and curate several synthetic and real-world benchmarks including analysis of human reviews and diagnosis of neuropathic and brain tumors, to comprehensively evaluate COAT. Extensive empirical results confirm the effectiveness and reliability of COAT with significant improvements. © 2024 Neural information processing systems foundation. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Spatio-Temporal Diffusion Model for Missing and Real-Time Financial Data Inference 24

A Spatio-Temporal Diffusion Model for Missing and Real-Time ...

引用

33rd ACM International conference on Information and Knowledge Management (CIKM)

作者： Fang, Yupeng Liu, Ruirui Huang, Huichou Zhao, Peilin Wu, Qingyao South China Univ Technol Guangzhou Peoples R China Brunel Univ London Kings Coll London London England City Univ Hong Kong Bayescien Technol Hong Kong Peoples R China Tencent AI Lab Shenzhen Peoples R China Peng Cheng Lab Shenzhen Peoples R China

ISBN: (纸本)9798400704369

Missing values and unreleased figures are common but highly important for backtesting and real-time analysis in the financial industry, yet underexploited in the existing literature. In this paper, we focus on the issue of empirical asset pricing, where the cross-section of future asset returns is a function of lagged firm characteristics that vary in time frequencies and missing ratios. Most of the existing imputation methods cannot fully capture the complex and evolving spatio-temporal relations among firm-level characteristics. In particular, these methods fail to explicitly consider the spatial relations and feature structure in the stock network where we have to process granular data of thousands of stocks and hundreds of characteristics for each stock. To address these challenges, we propose a spatio-temporal diffusion model (STDM) that gradually recovers the masked financial data conditioning on high-dimensional stock-and-characteristics historical data. We propose characteristic-specific projection to construct characteristic-level features at both ends of the STDM, meanwhile maintaining firm-level features in the middle of the STDM to largely reduce the computational memory. Moreover, along with the temporal attention, we design a spatial graph convolutional network, making it computationally efficient and effective to learn time-varying spatio-temporal interdependence across firms. We further employ an implicit sampler that greatly accelerates the inference procedure so that the STDM is able to produce high-quality point and density estimates of missing and real-time firm characteristics within a few steps. We evaluate our model on the most comprehensive open-source dataset 'OSAP' and generate state-of-the-art performance in extensive experiments.

关键词： Missing Value Imputation Real-Time Nowcasting Diffusion Model Financial Data processing

来源：评论

学校读者我要写书评

暂无评论

From Understanding to Utilization: A Survey on Explainability for Large language Models

arXiv

引用

arXiv 2024年

作者： Luo, Haoyan Specia, Lucia Imperial College London United Kingdom

Explainability for Large language Models (LLMs) is a critical yet challenging aspect of natural language processing. As LLMs are increasingly integral to diverse applications, their "black-box" nature sparks significant concerns regarding transparency and ethical use. This survey underscores the imperative for increased explainability in LLMs, delving into both the research on explainability and the various methodologies and tasks that utilize an understanding of these models. Our focus is primarily on pre-trained Transformer-based LLMs, such as LLaMA (Touvron et al., 2023), which pose distinctive interpretability challenges due to their scale and complexity. In terms of existing methods, we classify them into local and global analyses, based on their explanatory objectives. When considering the utilization of explainability, we explore several compelling methods that concentrate on model editing, control generation, and model enhancement. Additionally, we examine representative evaluation metrics and datasets, elucidating their advantages and limitations. Our goal is to reconcile theoretical and empirical understanding with practical implementation, proposing exciting avenues for explanatory techniques and their applications in the LLMs era. Copyright © 2024, The Authors. All rights reserved.

关键词： natural language processing systems

来源：评论

学校读者我要写书评

暂无评论

Unbiased Multimodal Audio-to-Intent Recognition

Unbiased Multimodal Audio-to-Intent Recognition

引用

International conference on Acoustics, Speech, and Signal processing (ICASSP)

作者： Qian Dong Yuezhou Dong Ke Qin Guiduo Duan Tao He University of Electronic Science and Technology of China Chengdu China

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Audio-to-intent recognition is a critical task focused on identifying a speaker’s intent from spoken language. Recently, multimodal audio-to-intent approaches have emerged as the predominant strategy for enhancing audio-based intent recognition. In this study, we conduct a series of empirical experiments that reveal a significant modality bias in current multimodal audio-to-intent recognition methods. Specifically, these methods disproportionately rely on the textual modality to determine intent, often neglecting the audio data. To address this issue, we propose a context-enhanced contrastive learning framework designed to capture rich regional- and global-audio context information, thereby enabling more balanced audio-to-intent recognition. Additionally, we introduce a prototype-based intent classification strategy that encourages different intent classes and modalities to converge toward unified prototypes, leading to smoother classification boundaries as opposed to the traditionally skewed boundaries. Extensive experiments demonstrate that our approach effectively mitigates modality bias, e.g., a performance improvement of 2.12% in intent classification compared to the state-of-the-art method GZAIR on the dataset MintRec.

关键词： Visualization Accuracy Intent recognition Prototypes Contrastive learning Signal processing Acoustics Speech processing

来源：评论

学校读者我要写书评

暂无评论

Analysis of Extractive Text Summarization methods as a Binary Classification Problem 1

引用

16th International conference on Computational Intelligence in Security for Information Systems, CISIS 2023 and 14th International conference on EUropean Transnational Education, ICEUTE 2023

作者： Komorniczak, Joanna Wojciechowski, Szymon Klikowski, Jakub Kozik, Rafal Choraś, Michal Pattern Recognition Team of Department of Computer Systems and Networks Wroclaw University of Science and Technology Wybrzeże Wyspiańskiego 27 Wroclaw50-370 Poland Bydgoszcz University of Science and Technology Al. prof. S. Kaliskiego 7 Bydgoszcz85-796 Poland

ISBN: (数字)9783031425196

ISBN: (纸本)9783031425189

One of the critical challenges for natural language processing methods is the issue of automatic content summarization. The enormous increase in the amount of data delivered to users by news services leads to an overload of information without meaningful content. There is a need to generate an automatic text summary that contains as much essential information as possible while keeping the resulting text smooth and concise. methods of automatic content summarization fall into two categories: extractive and abstractive. This work converts the task of extractive summarization to a binary classification problem. The research focused on analyzing various techniques for extracting the abstract in a supervised learning manner. The results suggest that this different view of text summarization has excellent potential. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

关键词： Text processing

来源：评论

学校读者我要写书评

暂无评论

Implementing Data Science and Big Data Analytics to Their Full Potential: Protocols for Gaining Useful Understanding and Promoting Data-Driven Innovation 5

Implementing Data Science and Big Data Analytics to Their Fu...

引用

5th IEEE Global conference for Advancement in Technology, GCAT 2024

作者： Arthy, M. Nivethitha, S.B. Karuppiah, S.P. Vijayanandh, T. Sumalatha, A. Ramesh, T. School of Computing College of Engineering and Technology Department of Data Science and Business Systems Tamil Nadu Kattankulathur India Rajalakshmi Engineering College Mechatronics Engineering Tamil Nadu Chennai India St.Joseph's College of Engineering Department of Mba India Vel Tech Rangarajan Dr. Sagunthala R&d Institute of Science and Technology School of Computing Department of Computer Science & Engineering Tamil Nadu Chennai India Kristu Jayanti College Autonomous Department of Computer Science Bengaluru India R.M.K Engineering College Department of Computer Science & Engineering Tamil Nadu Chennai India

ISBN: (纸本)9798350376685

The proper implementation of data science and big data analytics has become an essential component for gaining valuable insight and driving innovation across a variety of industries in this era, which is characterized by the constant presence of data. The strategies and protocols that are necessary to fully exploit the promise of these transformational technologies are investigated in depth by this research. A complete framework for enterprises to harness their data assets is the goal of the study, which will be accomplished by analyzing the fundamental principles of data science as well as the difficulties connected with big data. A number of advanced analytics methods, including machine learning, predictive modeling, and natural language processing, are being investigated as potential integrations into the decision-making process as part of this research. The relevance of data governance, quality, and the ethical considerations that surround data usage are also discussed in this document. It is demonstrated in this study that data-driven insights may lead to strategic advantages, operational efficiencies, and the development of creative goods and services. This is accomplished through a number of case studies and practical examples. In order for businesses to make the most of the potential that are given by data science and big data analytics, it is vital for them to have a culture of data literacy as well as a robust infrastructure, according to the results made. At the end of the study, a collection of best practices and recommendations for businesses and governments to implement in order to cultivate a data-driven environment that encourages innovation and usage of data in a responsible manner are presented. In light of the findings of this study, it is clear that data science and big data analytics require a more comprehensive approach, one that strikes a balance between technological capabilities, organizational readiness, and the influence on society. © 2024

关键词： Big Data Analytics Data Governance Data Infrastructure Data Literacy Data Quality Data Science Data-Driven Decision Making Data-Driven Innovation Ethical Considerations Machine Learning natural language processing Predictive Modeling Strategic Advantage

来源：评论

学校读者我要写书评

暂无评论

Research on Tibetan Verb Character Recognition Based on Bert Model

Research on Tibetan Verb Character Recognition Based on Bert...

引用

Artificial Intelligence and Computer Engineering (ICAICE), International conference on

作者： Ruohan Tian Dunzhu Laba Peilin Song Ziheng Xu School of Information Science and Technology Tibet University Lhasa Tibet Autonomous Region

ISBN: (数字)9798331528911

ISBN: (纸本)9798331528928

The traditional recognition methods of Tibetan verbs are mainly based on rule matching and statistical features,however,these methods are often limited by the complex structure of Tibetan language, rules are not fixed and other factors,the accuracy rate is low and the generalization ability is limited. The two-way information of the input text is taken into account in the coding process of Bert model, which avoids the loss of one-way information in the traditional model. By the Bert model,we have realized the accurate recognition of Tibetan verbs. The experimental results show that our method has a significant improvement on the recognition of verb error character, the accuracy and generalization ability are significantly improved compared with the traditional methods. The results of this study provide new ideas and methods for the research and application of Tibetan text processing and contribute to the development of Tibetan natural language processing technology.

关键词： Deep learning Adaptation models Accuracy Text recognition Computational modeling natural language processing Character recognition Optimization Context modeling Text processing

来源：评论

学校读者我要写书评

暂无评论

Automated Relation Extraction for Improved Generalizability across Different Types of Text

Automated Relation Extraction for Improved Generalizability ...

引用

ASCE International conference on Computing in Civil Engineering (I3CE)

作者： Chen, Qiyang El-Gohary, Nora Univ Illinois Dept Civil & Environm Engn Urbana IL 61801 USA

ISBN: (数字)9780784485231

ISBN: (纸本)9780784485231

Bridge textual reports capture technically detailed data/information about bridge conditions and maintenance actions, which offers opportunities to improve the prediction of future bridge conditions for further bridge maintenance decision-making support. To automatically analyze these reports, there is a need for relation extraction methods to extract relation information from the reports for linking recognized entities with predefined semantic categories (e.g., caused by) and representing the extracted semantic relations in a structured way. To address this need, this paper proposes a deep learning-based relation extraction model. The proposed model utilizes convolutional neural network (CNN) to encode sentence-level features, and bidirectional long short-term memory (BiLSTM) to build a relation extractor to capture the patterns of the predefined relation types. The proposed model was evaluated in extracting relations from multiple types of bridge-related textual reports for representing bridge defect information-including relations among bridge entities-in a semantically rich structured way.

关键词： Artificial intelligence natural language processing Information extraction Relation extraction Deep learning Bridge inspection Bridge conditions.

来源：评论

学校读者我要写书评

暂无评论

Block Diagram-to-Text: Understanding Block Diagram Images by Generating natural language Descriptors 2

Block Diagram-to-Text: Understanding Block Diagram Images by...

引用

2nd conference of the Asia-Pacific Chapter of the Association for Computational Linguistics and the 12th International Joint conference on natural language processing, AACL-IJCNLP 2022

作者： Bhushan, Shreyanshu Lee, Minho Department of Artificial Intelligence Kyungpook National University Korea Republic of Ali Co. Ltd. Korea Republic of

ISBN: (纸本)9781959429043

Block diagrams are very popular for representing a workflow or process of a model. Understanding block diagrams by generating summaries can be extremely useful in documentsummarization. It can also assist people in inferring key insights from block diagrams without requiring a lot of perceptual and cognitiveeffort. In this paper, we propose a novel taskof converting block diagram images into textby presenting a framework called "BloSum".This framework extracts the contextual meaning from the images in the form of triplets thathelp the language model in summary generation. We also introduce a new dataset for complex computerized block diagrams, explain thedataset preparation process, and later analyzeit. Additionally, to showcase the generalizationof the model, we test our method with publiclyavailable handwritten block diagram *** evaluation with different metrics demonstrates the effectiveness of our approach thatoutperforms other methods and techniques. © AACL-IJCNLP *** rights reserved

关键词：

来源：评论

学校读者我要写书评

暂无评论

Resolving duplicates in Large Multiple-Choice Questions Repositories 20

Resolving duplicates in Large Multiple-Choice Questions Repo...

引用

20th conference on Information and Research Science Connecting to Digital and Library Science, IRCDL 2024

作者： Albano, Valentina Firmani, Donatella Laura, Luigi Mathew, Jerin George Paoletti, Anna Lucia Torrente, Irene Dip. Funzione Pubblica Corso Vittorio Emanuele II 116 Rome00186 Italy Sapienza University Piazzale Aldo Moro 5 Rome00185 Italy Uninettuno University Corso Vittorio Emanuele II 39 Rome00186 Italy Formez Viale Marx 15 Rome00137 Italy

Multiple-choice questions (MCQs) are commonly used in educational assessments and professional certification examinations. However, managing vast collections of MCQs presents numerous challenges, including maintaining their quality and relevance. A notable issue in such repositories is the occurrence of conceptually identical questions presented in varied forms. These duplicates, while different in wording, fail to enhance the value of the repository. In this extended abstract, we present our approach for identifying and handling potential duplicate questions in large MCQ databases. Our proposed method involves three primary stages: initial pre-processing of MCQs, calculation of similarity based on natural language processing (NLP) techniques, and a graph-based method for exploring these similarities. © 2021 Copyright for this paper by its authors.

关键词： Graphic methods

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：