检索结果-内蒙古大学图书馆

arXiv 2023年

作者： Gurrapu, Sai Kulkarni, Ajay Huang, Lifu Lourentzou, Ismini Freeman, Laura Batarseh, Feras A. Department of Computer Science Virginia Tech United States Department of Computational and Data Sciences George Mason University United States Department of Statistics Virginia Tech United States Department of Electrical and Computer Engineering Virginia Tech United States

Recent advances in deep learning have improved the performance of many Natural Language Processing (NLP) tasks such as translation, question-answering, and text classification. However, this improvement comes at the expense of model explainability. Black-box models make it difficult to understand the internals of a system and the process it takes to arrive at an output. Numerical (LIME, Shapley) and visualization (saliency heatmap) explainability techniques are helpful;however, they are insufficient because they require specialized knowledge. These factors led rationalization to emerge as a more accessible explainable technique in NLP. Rationalization justifies a model’s output by providing a natural language explanation (rationale). Recent improvements in natural language generation have made rationalization an attractive technique because it is intuitive, human-comprehensible, and accessible to non-technical users. Since rationalization is a relatively new field, it is disorganized. As the first survey, rationalization literature in NLP from 2007-2022 is analyzed. This survey presents available methods, explainable evaluations, code, and datasets used across various NLP tasks that use rationalization. Further, a new subfield in Explainable AI (XAI), namely, Rational AI (RAI), is introduced to advance the current state of rationalization. A discussion on observed insights, challenges, and future directions is provided to point to promising research opportunities. © 2023, CC BY-SA.

关键词： Deep learning

来源：评论

学校读者我要写书评

暂无评论

On computing probabilistic explanations for decision trees 22

On computing probabilistic explanations for decision trees

引用

Proceedings of the 36th International Conference on Neural Information Processing Systems

作者： Marcelo Arenas Pablo Barceló Miguel Romero Bernardo Subercaseaux Department of Computer Science PUC Chile and Institute for Mathematical and Computational Engineering PUC Chile and Millenium Institute for Foundational Research on Data Chile Institute for Mathematical and Computational Engineering PUC Chile and Millenium Institute for Foundational Research on Data Chile and CENIA Chile Faculty of Engineering and Science UAI Chile and CENIA Chile Carnegie Mellon University Pittsburgh

ISBN: (纸本)9781713871088

Formal XAI (explainable AI) is a growing area that focuses on computing explanations with mathematical guarantees for the decisions made by ML models. Inside formal XAI, one of the most studied cases is that of explaining the choices taken by decision trees, as they are traditionally deemed as one of the most interpretable classes of models. Recent work has focused on studying the computation of sufficient reasons, a kind of explanation in which given a decision tree T and an instance x, one explains the decision T(x) by providing a subset y of the features of x such that for any other instance z compatible with y, it holds that T (z) = T (x), intuitively meaning that the features in y are already enough to fully justify the classification of x by T. It has been argued, however, that sufficient reasons constitute a restrictive notion of explanation. For such a reason, the community has started to study their probabilistic counterpart, in which one requires that the probability of T(z) = T(x) must be at least some value δ ∈ (0, 1], where z is a random instance that is compatible with y. Our paper settles the computational complexity of δ-sufficient-reasons over decision trees, showing that both (1) finding δ-sufficient-reasons that are minimal in size, and (2) finding δ-sufficient-reasons that are minimal inclusion-wise, do not admit polynomial-time algorithms (unless PTIME = NP). This is in stark contrast with the deterministic case (δ =1) where inclusion-wise minimal sufficient-reasons are easy to compute. By doing this, we answer two open problems originally raised by Izza et al., and extend the hardness of explanations for Boolean circuits presented by Wäldchen et al. to the more restricted case of decision trees. On the positive side, we identify structural restrictions of decision trees that make the problem tractable, and show how SAT solvers might be able to tackle these problems in practical settings.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Multi-Granular and Multi-Criteria Large-Scale Group Decision-Making Method Using the Aggressive Ordered Aggregation Operator Created Using Sentiment Analysis

SSRN

引用

SSRN 2022年

作者： Morfeq, Ali Trillo, José Ramón Morente-Molinera, Juan Antonio Al-Hmouz, Rami Herrera-Viedma, Enrique Department of Electrical and Computer Engineering The Faculty of Engineering The King Abdulaziz University Jeddah21589 Saudi Arabia Department of Computer Science and Artificial Intelligence Andalusian Research Institute in Data Science and Computational Intelligence DaSCI University of Granada Granada18071 Spain

Currently, human beings make different decisions daily, from choosing where to invest their money to decide how they can improve their business. Helping make the decision, Decision-Making processes can be used, in which several experts can participate. Nevertheless, when you want to make a Decision-Making process with large number of experts, Large-Scale Group Decision-Making methods are used, but several problems arise when using them. The first problem is how to manage all the information that is generated in the Group Decision-Making process, as the number of participants is larger than in a previous Group Decision-Making process. The second problem is knowing how to control the aggressiveness that the debate can generate when experts give their information. The third problem is the management of relationships in the debate, as differences can generate controversy and arguments. Solving these problems, this paper proposes a novel Large-Scale Group Decision-Making process applied to a multi-criteria environment and has a multi-granular fuzzy linguistic modelling that uses sentiment analysis to aggregate the criteria and optimises the consensus value. With this method, the experts can express themselves as they wish and it increases the relevance of the debate, as extra information is extracted and used during the process. © 2022, The Authors. All rights reserved.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Combining Self-Supervised Learning and Adversarial Training Based Domain Adaptation for Speaker Verification

Combining Self-Supervised Learning and Adversarial Training ...

引用

International Symposium on Chinese Spoken Language Processing

作者： Zhengyang Chen Shuai Wang Bing Han Yanmin Qian Department of Computer Science and Engineering Auditory Cognition and Computational Acoustics Lab MoE Key Lab of Artificial Intelligence AI Institute Shanghai Jiao Tong University Shanghai Shenzhen Research Institute of Big Data Shenzhen Chinese University of Hong Kong Shenzhen

ISBN: (数字)9798331516826

ISBN: (纸本)9798331516833

Adapting an existing well-trained system to a new domain using only unlabeled data is a highly sought-after yet challenging task for speaker verification in real-world scenarios. In this paper, we study two different domain adaptation methods, the adversarial domain adaptation (ADA) and the self-supervised learning-based domain adaptation (SSDA). To facilitate the deployment of unsupervised adaptation methods in applications, we conduct a detailed analysis of the characteristics of both the ADA and SSDA adaptation strategies. Our findings indicate that the SSDA strategy's performance is highly influenced by the amount of target domain data, whereas the ADA strategy is relatively insensitive to data quantity. Furthermore, augmenting target domain data enhances SSDA system performance but diminishes ADA performance. To further enhance system performance, we explore the complementarity between ADA and SSDA. Our results demonstrate that ADA and SSDA complement each other. When both strategies are applied jointly, the best system achieves over 20.0% relative Equal Error Rate (EER) improvement on the Cnceleb evaluation set and over 35.0% relative average EER improvement on the SRE16 Cantonese and Tagalog evaluation set under domain mismatched conditions.

关键词： Training Error analysis System performance Self-supervised learning Speech processing

来源：评论

学校读者我要写书评

暂无评论

Quantitative Analysis of IITs’ Research Growth and SDG Contributions

arXiv

引用

arXiv 2024年

作者： Sharma, Kiran Nagori, Akshat Manya Dubey, Mehul Khurana, Parul School of Engineering & Technology BML Munjal University Haryana Gurugram122413 India Center for Advanced Data and Computational Science BML Munjal University Haryana Gurugram122413 India School of Computer Applications Lovely Professional University Punjab Phagwara144411 India

The Indian Institutes of Technology (IITs) are vital to India’s research ecosystem, advancing technology and engineering for industrial and societal benefits. This study reviews the research performance of top IITs—Bombay, Delhi, Madras, Kharagpur, and Kanpur based on Scopus-indexed publications (1952–2024). Research output has grown exponentially, supported by increased funding and collaborations. IIT-Kanpur excels in research impact, while IIT-Bombay and IIT-Madras are highly productive but show slightly lower per-paper impact. Internationally, IITs collaborate robustly with the USA, Germany, and the UK, alongside Asian nations like Japan and South Korea, with IIT-Madras leading inter-IIT partnerships. Research priorities align with SDG 3 (Health), SDG 7 (Clean Energy), and SDG 11 (Sustainable Cities). Despite strengths in fields like energy, fluid dynamics, and materials science, challenges persist, including limited collaboration with newer IITs and gaps in emerging fields. Strengthening specialization and partnerships is crucial for addressing global challenges and advancing sustainable development. © 2024, CC BY.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Validating the information technology (IT) implementation framework to Implement mHealth technology for consumers: A case study of the Sense2Quit app for smoking cessation

引用

International Journal of Medical Informatics 2025年 202卷 105977-105977页

作者： Brin, Maeve Fontalvo, Sydney Hu, David Cioe, Patricia Huang, Ming-Chun Xu, Wenyao Schnall, Rebecca Columbia University School of Nursing New YorkNY United States Brown University School of Public Health United States Department of Data and Computational Science Duke Kunshan University Jiangsu China Department of Computer Science & Engineering University at Buffalo the State University of New York BuffaloNY United States Columbia University School of Public Health New YorkNY United States

Objective: The goal of this paper was to understand the applicability of the Information Technology (IT) Implementation Framework, a multi-level approach to identify factors that impede or promote IT usage, for incorporating a mHealth technology for consumers in the community setting. Methods: A case study of the implementation of the Sense2Quit App for smoking cessation among people living with HIV was examined to parse out the factors within the framework that are applicable to mHealth technology and the factors that may need modification for use of this framework within this context. Results: Findings suggest that phases two through five of the framework were applicable to our study and phase one was not. Conclusion: Findings support the use of the theory for implementation of mHealth technology for promoting consumer health at the community level. This use case may be useful for stakeholders evaluating implementation of mHealth for patients with chronic conditions as it highlights the need to identify preferences of app specifications, personal habits, and various factors such as confidentiality and digital literacy which may challenge sustained usage. © 2025 Elsevier B.V.

关键词： mHealth

来源：评论

学校读者我要写书评

暂无评论

Improved Classification of Alzheimer’s Disease With Convolutional Neural Networks

Improved Classification of Alzheimer’s Disease With Convolu...

引用

IEEE Signal Processing in Medicine and Biology Symposium (SPMB)

作者： S. Muhammed J. Upadhya S. Poudel M. Hasan K. Donthula J. Vargas J. Ranganathan K. Poudel Computational and Data Science Middle Tennessee State University Murfreesboro TN USA Computer Science Middle Tennessee State University Murfreesboro TN USA Optum Inc Murfreesboro TN USA Engineering Technology Middle Tennessee State University Murfreesboro TN USA

This work focuses on classifying MRI images using machine learning models to identify Alzheimer’s disease (AD), the most common form of dementia, at an early stage [1]. It is now possible to identify and forecast the onset of AD by analyzing brain scans collected through Magnetic Resonance Imaging (MRI) and using artificial intelligence (AI) technologies, classifying patients as either at risk or not. The main goal is to make precise predictions. Improved prediction and detection tools for radiologists, physicians, and caregivers will be made available as a result of the study’s determination of the likelihood that individuals would develop AD and proper categorization of them. Using machine learning methods such as Convolutional Neural Network (CNN), Support Vector Machine (SVM), and fastai—a user-friendly deep learning library and framework—we developed models using a dataset of 6400 MRI images from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) 4 class for the early detection and classification of AD [2]. The models achieve an impressive accuracy rate of 84.4%, with the addition of a pre-trained Visual Geometry Group (VGG) layer to the sequential model, outperforming other algorithms. This study demonstrates a viable method for classifying and diagnosing Alzheimer’s disease early on, utilizing MRI images and machine learning models. The acquired accuracy rates show the models’ potential to assist doctors and other caregivers in AD diagnosis and management, thereby maximizing overall effectiveness.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Predicting the Cerebral Blood Flow Change Condition during Brain Strokes using Feature Fusion of FMRI Images and Clinical Features

Predicting the Cerebral Blood Flow Change Condition during B...

引用

International Conference on Computing and Networking Technology (ICCNT)

作者： Vandana Sharma Anurag Sinha Michael Wiryaseputra Biresh Kumar Tarun Raj Kumar Ahmed Alkhayyat Institute of Information Technology Christ University Delhi Campus India Department of Computer Science and Information Technology IGNOU New Delhi India Soegijapranata Catholic University Semarang Indonesia Department of Computer Science Amity Univeristy Jharkhand India Computational Mathematics & Data Science IIT Madras India College of Technical Engineering The Islamic University Najaf Iraq

By fusing clinical information with functional magnetic resonance imaging (fFMRI) pictures, this study describes a novel method for predicting changes in cerebral blood flow during brain strokes. The FMRI data and patient-specific variables, such as age, gender, and medical history, are combined via feature fusion in the proposed technique. As a result, the model developed can accurately forecast changes in cerebral blood flow that occur during brain strokes. The efficiency of the suggested strategy is shown by experimental findings. The performance of the model is greatly enhanced when FMRI data and clinical characteristics are combined as opposed to just one data source. The findings of this study have important ramifications for increasing the accuracy of stroke diagnosis and treatment and, eventually, for bettering patient outcomes. The experimental results showed that the proposed method a high level of accuracy in predicting changes in cerebral blood flow after brain strokes. The performance of the model was much enhanced by combining clinical characteristics with FMRI data as opposed to using only one of these data sources. This emphasizes the value of including pertinent clinical information in the diagnosis and management of stroke.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Decoding Speech Categorization using Microstate Cortical EEG Signals and Machine Learning

Decoding Speech Categorization using Microstate Cortical EEG...

引用

IEEE Signal Processing in Medicine and Biology Symposium (SPMB)

作者： M. Mahmud M. Hasan M. Yeasin G. Bidelman Div. of General Internal Medicine Univ. of Tennessee Health Science Center Memphis TN USA Department of Electrical and Computer Engineering University of Memphis TN USA Computational and Data Science Middle Tennessee State University Murfreesboro TN USA Department of Speech Language and Hearing Sciences Indiana University Bloomington IN USA

ISBN: (数字)9798350388572

ISBN: (纸本)9798350388589

Categorical perception (CP) is a perceptual phenomenon that refers to the tendency of humans to group speech sounds into discrete units. In this work, we used cortical event-related potential (ERP) signals, recorded during a speech identification task, as input to a Hierarchical Dirichlet Process Hidden Markov Model (HDP-HMM) and machine learning (ML) classifiers to examine how neural signatures of the brain distinguish prototypical vs. ambiguous speech token categories. Particularly, we used extreme gradient boosting (XGBoost), support vector machine, and random forest classifiers. Our analysis shows that using source reconstructed whole-brain data, the XGBoost classifier yielded the best classification accuracy of 94.12% with (area under the curve (AUC) 94.13%, Precision 94.00%, F1-score 94.00%) in the 197-258 ms time window after the stimulus onset. We also identified 15 important brain regions that distinguished vowel classes with accuracy of 90.28% (AUC 90.17%, F1-score 90.00%, Precision 90.00%), 4% less than whole-brain data. Remarkably, out of these 15 critical brain regions, nine were from the left hemisphere, consistent with the left-brain dominance for language processing. The best speech decoding was obtained from early portions of the ERP (~180-320 ms), which corroborates notions that sensory encoding is critical for successful phonetic processing.

关键词： Support vector machines Accuracy Hidden Markov models Phonetics Signal processing Electroencephalography Real-time systems Decoding Speech processing Random forests

来源：评论

学校读者我要写书评

暂无评论

Revolutionizing Image Analysis: The Integration of MLP and GRU Architectures in Advanced computer Vision 1

Revolutionizing Image Analysis: The Integration of MLP and G...

引用

1st IEEE International Conference on Green Industrial Electronics and Sustainable Technologies, GIEST 2024

作者： Nimma, Divya Sengar, Alok Singh Kaur, Chamandeep Khan, Shamim Ahmad Anushna, S.D. Raj, I Infant University of Southern Mississippi Data Analyst in UMMC PhD in Computational Science United States Noida International University School of Sciences Dept. of Computer Science Uttar Pradesh Greater Noida India Jazan University Department of Computer Science Jazan Saudi Arabia Glocal University Glocal School of Science & Technology U.P India Hyderabad India K.Ramakrishnan College of Engineering Department of CSE Tamil Nadu Trichy India

ISBN: (纸本)9798331540142

Although there have been significant advances in computer vision recently, there are still difficulties in efficiently interpreting dynamic and complicated visual input. Through the integration of Gated Recurrent Units (GRUs) and Multi-Layer Perceptrons (MLPs) in advanced picture analysis, this study suggests a novel way to overcome these issues. Utilizing the advantages of both systems, the integrated model seeks to improve visual information comprehension and interpretation. computer vision applications involving temporal context and semantic understanding are sometimes difficult for traditional approaches to handle. MLPs are good at collecting spatial features;on the other hand, GRUs are efficient at processing data sequentially. This research aims to develop and evaluate an integrated MLP-GRU model for advanced image analysis, combining the strengths of MLPs and GRUs. The model aims to achieve superior performance in tasks like object detection, image classification, and semantic segmentation. The model combines MLPs for spatial feature extraction and GRUs for capturing temporal dependencies within sequential data. The experimental results show significant improvements in accuracy of 98%, compared to traditional approaches. The model also demonstrates the ability to discern complex visual patterns and make accurate predictions. This study highlights the potential of hybrid neural network architectures in image analysis, highlighting the paradigm shift in computer vision and their ability to analyze diverse and dynamic visual data. © 2024 IEEE.

关键词： Semantic Segmentation

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：