检索结果-内蒙古大学图书馆

3rd International Conference on Emerging Electronics and Automation, E2A 2023

作者： Rao, S. V. Achuta Murty, P. T. Satyanarayana Maram, Balajee Garg, Ankit Daniya, T. Kumar, B. Santhosh Data Science Research Laboratories Sree Dattha Institute of Engineering and Science Sheriguda Telangana Hyderabad501 510 India Shri Vishnu Engineering College for Women Bhimavaram India AIT- Department of Computer Science and Engineering Chandigarh University Punjab Gharuan Mohali140055 India GMR Institute of Technology Rajam532127 India Telangana Ibrahimpatnam501506 India

ISBN: (纸本)9789819768011

Accurate and prompt diagnosis is crucial to the treatment of patients with dermatological diseases, which are a significant cause for worry in the medical community. By automating the diagnosing process, deep learning algorithms have recently demonstrated significant promise in the field of dermatology. In-depth research on the use of deep transfer learning for multiclass dermatological classification is presented in this publication. To train deep neural networks, the study makes use of a broad and comprehensive dermatology dataset that includes a variety of skin disorders. Modern pre-trained models are used with transfer learning techniques to make it easier to extract discriminative features. In order to improve classification performance, we delve into the complex process of model fine-tuning and investigate data augmentation techniques. Our research shows that multiclass dermatology classification accuracy has significantly improved, with models delivering reliable findings for a variety of skin disorders. We offer a thorough evaluation of the model's performance, emphasizing its ability to distinguish between disorders that appear to be visually identical and assisting dermatologists in their diagnostic procedures. The field of dermatology will be significantly impacted by the study's findings. Our work helps to the creation of precise, effective, and available dermatological diagnostic tools by bridging the gap between conventional diagnosis techniques and cutting-edge deep learning technologies. We also go over potential difficulties, ethical issues, and prospective future research routes, highlighting how crucial it is to keep this field innovative. This thorough study highlights the importance of artificial intelligence in improving patient care in the field of dermatology and the possibility of deep transfer learning to revolutionize dermatology diagnosis. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Medical imaging

来源：评论

学校读者我要写书评

暂无评论

Not all distributional shifts are equal: Fine-grained robust conformal inference

arXiv

引用

arXiv 2024年

作者： Ai, Jiahao Ren, Zhimei School of Mathematical Sciences Peking University China Department of Statistics and Data Science University of Pennsylvania United States

We introduce a fine-grained framework for uncertainty quantification of predictive models under distributional shifts. This framework distinguishes the shift in covariate distributions from that in the conditional relationship between the outcome (Y ) and the covariates (X). We propose to reweight the training samples to adjust for an identifiable covariate shift while protecting against worst-case conditional distribution shift bounded in an f-divergence ball. Based on ideas from conformal inference and distributionally robust learning, we present an algorithm that outputs (approximately) valid and efficient prediction intervals in the presence of distributional shifts. As a use case, we apply the framework to sensitivity analysis of individual treatment effects with hidden confounding. The proposed methods are evaluated in simulation studies and three real data applications, demonstrating superior robustness and efficiency compared with existing benchmarks. Copyright © 2024, The Authors. All rights reserved.

关键词： Sensitivity analysis

来源：评论

学校读者我要写书评

暂无评论

Finite-Sample Analysis of Learning High-Dimensional Single ReLU Neuron

arXiv

引用

arXiv 2023年

作者： Wu, Jingfeng Zou, Difan Chen, Zixiang Braverman, Vladimir Gu, Quanquan Kakade, Sham M. Department of Computer Science Johns Hopkins University United States Department of Computer Science The University of Hong Kong Hong Kong Department of Computer Science University of California Los Angeles United States Department of Computer Science Rice University United States Department of Computer Science Department of Statistics Harvard University United States

This paper considers the problem of learning a single ReLU neuron with squared loss (a.k.a., ReLU regression) in the overparameterized regime, where the input dimension can exceed the number of samples. We analyze a Perceptron-type algorithm called GLM-tron (Kakade et al., 2011) and provide its dimension-free risk upper bounds for high-dimensional ReLU regression in both well-specified and misspecified settings. Our risk bounds recover several existing results as special cases. Moreover, in the well-specified setting, we provide an instance-wise matching risk lower bound for GLM-tron. Our upper and lower risk bounds provide a sharp characterization of the high-dimensional ReLU regression problems that can be learned via GLM-tron. On the other hand, we provide some negative results for stochastic gradient descent (SGD) for ReLU regression with symmetric Bernoulli data: if the model is well-specified, the excess risk of SGD is provably no better than that of GLM-tron ignoring constant factors, for each problem instance;and in the noiseless case, GLM-tron can achieve a small risk while SGD unavoidably suffers from a constant risk in expectation. These results together suggest that GLM-tron might be preferable to SGD for high-dimensional ReLU regression. Copyright © 2023, The Authors. All rights reserved.

关键词： Gradient methods

来源：评论

学校读者我要写书评

暂无评论

Conditional Adversarial Learning for Empathetic Dialogue Response Generation

引用

APSIPA Transactions on Signal and Information Processing 2023年第1期12卷

作者： Su, Ming-Hsiang Wu, Chung-Hsien Liao, Chia-Yu Department of Data Science Soochow University Taipei Taiwan Department of Computer Science and Information Engineering National Cheng Kung University Tainan Taiwan

This study presents a novel approach for generating empathetic responses in dialogue through conditional adversarial learning. The method involves using a BERT-MLP model to detect the user's emotions and the system's dialogue act, and then utilizing conditional adversarial learning to construct a generator based on the user's emotions, dialogue history, and dialogue act. A sympathy discriminator is trained to distinguish between empathetic and non-empathetic responses, and the corresponding words are filled in the generated template based on the semantic slots. To evaluate the proposed approach, the study collected 1,740 conversations with empathetic responses, which were labeled with the user's emotion, medical history, and system dialogue act. The experimental results based on 5-fold cross-validation showed that the proposed method of applying conditional adversarial learning achieved the best BLEU score (41.3%), the best BERTSCORE (−5.84 for the evaluation on question sentences and generated sentences;−4.15 for the evaluation on answer sentences and generated sentences) and emotion reflection rate (86.4%), which outperformed the Transformer- and conditional Transformer-based methods. This study also conducted subjective evaluations, achieving 77.55%, 79.47%, and 75.87% accuracy in the scores of relevance, grammatical correctness, and empathy, respectively. In significance test and Cohen's KAPA score of relevance, grammatical correctness and empathy, the proposed method was better than the Transformer- and conditional Transformer-based methods. In addition, in the consulting performance evaluation, the experimental results showed that the proposed method achieved the best empathy score of 3.8 (average KAPA score was 0.627), which was better than the other methods. © 2023 M.-H. Su, C.-H. Wu and C.-Y. Liao.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

SOrdered and colored subgraph density problems

arXiv

引用

arXiv 2024年

作者： Cairncross, Emily Mubayi, Dhruv Department of Mathematics Statistics and Computer Science University of Illinois ChicagoIL60607 United States

We consider three extremal problems about the number of copies of a fixed graph in another larger graph. First, we correct an error in a result of Reiher and Wagner [13] and prove that the number of k-edge stars in a graph with density x ∈ [0, 1] is asymptotically maximized by a clique and isolated vertices or its complement. Next, among ordered n-vertex graphs with m edges, we determine the maximum and minimum number of copies of a k-edge star whose nonleaf vertex is minimum among all vertices of the star. Finally, for s ≥ 2, we define a particular 3-edge-colored complete graph F on 2s vertices with colors blue, green and red, and determine, for each (xb, xg) with xb + xg ≤ 1 and xb, xg ≥ 0, the maximum density of F in a large graph whose blue, green and red edge sets have densities xb, xg and 1 − xb − xg, respectively. These are the first nontrivial examples of colored graphs for which such complete results are proved. © 2024, CC BY.

关键词： Stars

来源：评论

学校读者我要写书评

暂无评论

Study on Online Review Based Consumer sentimental Analysis using Machine Learning Approaches

Study on Online Review Based Consumer sentimental Analysis u...

引用

2022 IEEE World Conference on Applied Intelligence and Computing, AIC 2022

作者： Sharon Roji Priya, C. Deepalakshmi, P. Cmr Institute of Technology Department of Artificial Intelligence & Data Science Karnataka Bengaluru560037 India Kalasalingam Academy of Research and Education Department of Computer Science and Engineering Tamilnadu Krishnankoil626126 India

ISBN: (数字)9781665479882

ISBN: (纸本)9781509050017

Analysing a vast quantity of social media data, which expands itself in volume, subjectivity, and heterogeneity on a manual basis becomes more difficult as technology progresses. In real-world applications, machine learning techniques are being used to address this issue. The goal of this article is to describe research that was conducted to assess the utility, breadth, and application of machine learning algorithms for Consumer Sentiment Analysis (CSA) in online reviews. We present a thorough evaluation of the literature in order to evaluate, examine, study and understand methodologies with directions, in order to uncover research gaps, hence showing the pairing's potential reach in the future. The major purpose is to read and analyse machine learning techniques used in the hotel and tourist industry to analyse customer sentiment in online evaluations. This research is crucial for service providers since it enables them to design customer management strategies for service selection. Additionally, there is a significant influence on scholars' future study orientations. © 2022 IEEE.

关键词： Sentiment analysis

来源：评论

学校读者我要写书评

暂无评论

Leveraging Deep Learning Models for Machine-Generated Text Detection Using Transformer-Based Models

Leveraging Deep Learning Models for Machine-Generated Text D...

引用

Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES), International Conference on

作者： P. Suthanthiradevi G Revathy K. Rejini Muthu Lakshmi.V Department of Data Science and Business Systems School of Computing SRM institute of Science and Technology Kattankulathur Chennai Department of Computer science and engineering Vels Institute of science Technology and Advanced studies Chennai Department of Computer Science Engineering Amrita college of Engineering and Technology Nagercoil Department of Computer Science and Engineering St. Joseph's College of Engineering OMR Chennai

ISBN: (数字)9798331543624

ISBN: (纸本)9798331543631

GPT is a large language model (LLM) derived from natural language processing that can generate a human-like text using machine learning. However, these models raise questions about authenticity and reliability of material, particularly in fields such as journalism, social media, and academia, despite their usefulness for automating text-based tasks. Detecting machine-generated text is thus an important difficulty in ensuring content integrity. This study investigates the use of huge language models as a technique for recognizing machine-generated material. The author proposes a comprehensive detection model by evaluating the language patterns, syntactic structures, and stylistic traits that separate AI-generated literature from human writing. In addition, this research investigate the possibilities of fine-tuning models designed expressly for text identification tasks and evaluate their performance using LLM - Detect AI Generated Text datasets. In digital ecosystems, LLMs are effective at detecting AI-generated text, providing a novel approach for content moderation, academic integrity checks, and synthetic media detection. An increasingly AI-powered future will require a model that can discriminate between human and machine-generated writing in real-time. According to experimental findings, the CNN architecture's design combined with the use of DistilBERT embeddings allows for the effective and efficient classification of AI generated text data, achieving an exceptional 98% accuracy rate.

关键词： Deep learning Deepfakes Accuracy Social networking (online) Biological system modeling Large language models Text detection Writing Transformers Natural language processing

来源：评论

学校读者我要写书评

暂无评论

Real-Time Speech Emotion Recognition Using Hybrid Transformer Networks for Mental Health Monitoring

Real-Time Speech Emotion Recognition Using Hybrid Transforme...

引用

2025 International Conference on Multi-Agent Systems for Collaborative Intelligence, ICMSCI 2025

作者： Durga, P. Janani Basha, C. H. Hussaian Christopher, Nelson Kennedy Babu Senthilkumar, S. Kumar, N. Rathina Kavitha, M. E.G.S. Pillay Engineering College Department of Electronics and Communication Engineering Tamil Nadu Nagapattinam611002 India Sr University Department of Electrical and Electronics Engineering Telangana Hanumakonda50637 India Saveetha School of Engineering Saveetha Institute of Medical and Technical Sciences Department of Computer Science and Engineering Tamil Nadu Chennai602105 India Sns College of Engineering Department of Artificial Intelligence and Data Science Tamilnadu Coimbatore641107 India

ISBN: (纸本)9798331509828

This study proposes a hybrid method to multi dialect speech recognition using the transfer learning with transformer-based architecture. The primary objective of this method is to enhance the automatic speech recognition (ASR) across the different dialectal variations of specific language. The transfer learning methods are utilized to modify the base model that was trained on high resource language to manage the multiple dialects with limited labelled data for each dialect. This is achieved by using pre trained transformer methods such as Wav2Vec2.0 and Whisper. An approach which employs dialect specific embeddings is presented. It enables the method to capture data and differentiate between prosodic elements and phonetic that are distinctive to a variety of dialects. data augmentation technique are utilized in order to overcome the limited availability of labelled data for low resource dialects. These methods are utilized to increase the efficiency of training set and enhance the robustness of the method. Evaluation of the method is carried out using multi dialect speech datasets which include Arabic MGB-3 and common voice. These datasets represent a variety of dialects that are spoken in various linguistic regions. The result of this study indicates that the transformer-based method is effective in managing difficulties of dialectal variances while the reliability is reduced on different datasets. Different real-world applications are provided by the proposed method for improving voice enabled applications. Furthermore, this method leads the way for the development of ASR method that are more accurate and independent in locations that have a significant linguistic variation. © 2025 IEEE.

关键词： Transfer learning

来源：评论

学校读者我要写书评

暂无评论

A General Purpose Turkish CLIP Model (TrCLIP) for Image&Text Retrieval and its Application to E-Commerce 16

A General Purpose Turkish CLIP Model (TrCLIP) for Image&Text...

引用

16th International Conference on INnovations in Intelligent SysTems and Applications, INISTA 2022

作者： Ani, Yusuf Amasyali, Mehmet Fatih Hepsiburada Search-Data Science Team Istanbul Turkey Yildiz Technical University Computer Engineering Department Istanbul Turkey

ISBN: (数字)9781665498104

ISBN: (纸本)9781665498104

In this paper, we introduce a Turkish adaption of CLIP (Contrastive Language-Image Pre-Training). Our approach is to train a model with the same output space as the Text encoder of the CLIP model while processing Turkish input. For this, we collected 2.5M unique English-Turkish data. The model we named TrCLIP performed 71% in CIFAR100, 86% in VOC2007, and 47% in FER2013 as zero-shot accuracy. We have examined its performance on e-commerce data and a vast domain-independent dataset in image and text retrieval tasks. The model can work in Turkish without any extra fine-tuning. Models and dataset can be reachable from https://***/yusufani/TrCLIP. © 2022 IEEE.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

Population Age Group Sensitivity for COVID-19 Infections with Deep Learning

arXiv

引用

arXiv 2023年

作者： Islam, Md Khairul Valentine, Tyler Wang, Royal Davis, Levi Manner, Matt Fox, Judy Computer Science Department University of Virginia Charlottesville United States School of Data Science University of Virginia Charlottesville United States Computer Science Department School of Data Science University of Virginia Charlottesville United States

The COVID-19 pandemic has created unprecedented challenges for governments and healthcare systems worldwide, highlighting the critical importance of understanding the factors that contribute to virus transmission. This study aimed to identify the most influential age groups in COVID-19 infection rates at the US county level using the Modified Morris Method and deep learning for time series. Our approach involved training the state-of-the-art time-series model Temporal Fusion Transformer on different age groups as a static feature and the population vaccination status as the dynamic feature. We analyzed the impact of those age groups on COVID-19 infection rates by perturbing individual input features and ranked them based on their Morris sensitivity scores, which quantify their contribution to COVID-19 transmission rates. The findings are verified using ground truth data from the CDC and US Census, which provide the true infection rates for each age group. The results suggest that young adults were the most influential age group in COVID-19 transmission at the county level between March 1, 2020, and November 27, 2021. Using these results can inform public health policies and interventions, such as targeted vaccination strategies, to better control the spread of the virus. Our approach demonstrates the utility of feature sensitivity analysis in identifying critical factors contributing to COVID-19 transmission and can be applied in other public health domains. Copyright © 2023, The Authors. All rights reserved.

关键词： COVID-19

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：