检索结果-内蒙古大学图书馆

12th International Conference on Big data and Artificial Intelligence, BDA 2024

作者： Singh, Akansha Chakraborty, Soumendu Adhikari, Mainak Department of Computer Science Indian Institute of Information Technology Lucknow Lucknow India School of Data Science Indian Institute of Science Education and Research Thiruvananthapuram Thiruvananthapuram India

ISBN: (纸本)9783031818202

Air quality significantly impacts human health and economic conditions, making precise and timely assessment crucial in urban areas. Existing studies often fail to predict pollution accurately in smaller areas due to varying meteorological conditions. Most existing prototypes rely on standard ML models, trained with generic, publicly available datasets, which often lack local dynamics, adaptability, and synchronization. To address these challenges, in this paper, we propose a new Hybrid Edge-Cloud Air Quality Monitoring System. For initial model training a robust testbed is created by data Fusion of air quality data from diverse sources like Application Programming Interfaces and public datasets. The system trains and refines ML models in the cloud, identifies key features to reduce feature dimensions, and leverages local decision-making using site-specific data collected by sensors on edge devices. Sensors monitor local environmental parameters for real-time analysis and decision-making, ensuring accurate predictions for specific urban locations while optimizing performance and minimizing resource usage on edge devices. Explainable AI techniques improve transparency and confidence in decisions. The model is validated by simulating various ML models on the testbed, with and without feature extraction. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Prediction models

来源：评论

学校读者我要写书评

暂无评论

ZEROTH-ORDER OPTIMIZATION MEETS HUMAN FEEDBACK: PROVABLE LEARNING VIA RANKING ORACLES 12

ZEROTH-ORDER OPTIMIZATION MEETS HUMAN FEEDBACK: PROVABLE LEA...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Tang, Zhiwei Rybin, Dmitry Chang, Tsung-Hui School of Science and Engineering The Chinese University of Hong Kong Shenzhen China School of Data Science The Chinese University of Hong Kong Shenzhen China Shenzhen Research Institute of Big Data Shenzhen China

In this study, we delve into an emerging optimization challenge involving a black-box objective function that can only be gauged via a ranking oracle-a situation frequently encountered in real-world scenarios, especially when the function is evaluated by human judges. Such challenge is inspired from Reinforcement Learning with Human Feedback (RLHF), an approach recently employed to enhance the performance of Large Language Models (LLMs) using human guidance (Ouyang et al., 2022;Liu et al., 2023;OpenAI, 2022;Bai et al., 2022). We introduce ZO-RankSGD, an innovative zeroth-order optimization algorithm designed to tackle this optimization problem, accompanied by theoretical assurances. Our algorithm utilizes a novel rank-based random estimator to determine the descent direction and guarantees convergence to a stationary point. Moreover, ZO-RankSGD is readily applicable to policy optimization problems in Reinforcement Learning (RL), particularly when only ranking oracles for the episode reward are available. Last but not least, we demonstrate the effectiveness of ZO-RankSGD in a novel application: improving the quality of images generated by a diffusion generative model with human ranking feedback. Throughout experiments, we found that ZO-RankSGD can significantly enhance the detail of generated images with only a few rounds of human feedback. Overall, our work advances the field of zeroth-order optimization by addressing the problem of optimizing functions with only ranking feedback, and offers a new and effective approach for aligning Artificial Intelligence (AI) with human intentions. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

NEURON-ENHANCED AUTOENCODER MATRIX COMPLETION AND COLLABORATIVE FILTERING: THEORY AND PRACTICE 12

NEURON-ENHANCED AUTOENCODER MATRIX COMPLETION AND COLLABORAT...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Fan, Jicong Chen, Rui Zhang, Zhao Ding, Chris H.Q. School of Data Science The Chinese University of Hong Kong Shenzhen China Shenzhen Research Institute of Big Data Shenzhen China School of Computer Science & Information Engineering Hefei University of Technology China

Neural networks have shown promising performance in collaborative filtering and matrix completion but the theoretical analysis is limited and there is still room for improvement in terms of the accuracy of recovering missing values. This paper presents a neuron-enhanced autoencoder matrix completion (AEMC-NE) method and applies it to collaborative filtering. Our AEMC-NE adds an element-wise autoencoder to each output of the main autoencoder to enhance the reconstruction capability. Thus it can adaptively learn an activation function for the output layer to approximate possibly complicated response functions in real data. We provide theoretical analysis for AEMC-NE as well as AEMC to investigate the generalization ability of autoencoder and deep learning in matrix completion, considering both missing completely at random and missing not at random. We show that the element-wise neural network has the potential to reduce the generalization error bound, the data sparsity can be useful, and the prediction performance is closely related to the difference between the numbers of variables and samples. The numerical results on synthetic data and five benchmark datasets demonstrated the effectiveness of AEMC-NE in comparison to many baselines. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Collaborative filtering

来源：评论

学校读者我要写书评

暂无评论

GoMatching: A Simple Baseline for Video Text Spotting via Long and Short Term Matching 38

GoMatching: A Simple Baseline for Video Text Spotting via Lo...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： He, Haibin Ye, Maoyuan Zhang, Jing Liu, Juhua Du, Bo Tao, Dacheng School of Computer Science National Engineering Research Center for Multimedia Software Institute of Artificial Intelligence Wuhan University China College of Computing & Data Science Nanyang Technological University Singapore

Beyond the text detection and recognition tasks in image text spotting, video text spotting presents an augmented challenge with the inclusion of tracking. While advanced end-to-end trainable methods have shown commendable performance, the pursuit of multi-task optimization may pose the risk of producing sub-optimal outcomes for individual tasks. In this paper, we identify a main bottleneck in the state-of-the-art video text spotter: the limited recognition capability. In response to this issue, we propose to efficiently turn an off-the-shelf query-based image text spotter into a specialist on video and present a simple baseline termed GoMatching, which focuses the training efforts on tracking while maintaining strong recognition performance. To adapt the image text spotter to video datasets, we add a rescoring head to rescore each detected instance's confidence via efficient tuning, leading to a better tracking candidate pool. Additionally, we design a long-short term matching module, termed LST-Matcher, to enhance the spotter's tracking capability by integrating both long- and short-term matching results via Transformer. Based on the above simple designs, GoMatching delivers new records on ICDAR15-video, DSText, BOVText, and our proposed novel test set with arbitrary-shaped text termed ArTVideo, which demonstrates GoMatching's capability to accommodate general, dense, small, arbitrary-shaped, Chinese and English text scenarios while saving considerable training budgets. © 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

An In-depth Investigation of Sparse Rate Reduction in Transformer-like Models

arXiv

引用

arXiv 2024年

作者： Hu, Yunzhe Zou, Difan Xu, Dong School of Computing and Data Science The University of Hong Kong Hong Kong School of Computing and Data Science Institute of Data Science The University of Hong Kong Hong Kong

Deep neural networks have long been criticized for being black-box. To unveil the inner workings of modern neural architectures, a recent work [45] proposed an information-theoretic objective function called Sparse Rate Reduction (SRR) and interpreted its unrolled optimization as a Transformer-like model called Coding Rate Reduction Transformer (CRATE). However, the focus of the study was primarily on the basic implementation, and whether this objective is optimized in practice and its causal relationship to generalization remain elusive. Going beyond this study, we derive different implementations by analyzing layer-wise behaviors of CRATE, both theoretically and empirically. To reveal the predictive power of SRR on generalization, we collect a set of model variants induced by varied implementations and hyperparameters and evaluate SRR as a complexity measure based on its correlation with generalization. Surprisingly, we find out that SRR has a positive correlation coefficient and outperforms other baseline measures, such as path-norm and sharpness-based ones. Furthermore, we show that generalization can be improved using SRR as regularization on benchmark image classification datasets. We hope this paper can shed light on leveraging SRR to design principled models and study their generalization ability. © 2024, CC BY.

关键词：

来源：评论

学校读者我要写书评

暂无评论

WNet: Multi-Task Artistic Portrait Generation Method 14

WNet: Multi-Task Artistic Portrait Generation Method

引用

14th International Conference on Information Technology in Medicine and Education, ITME 2024

作者： Li, Juan Wu, Binlong Lin, Yuhui Zhang, Huang Huang, Xiangping Fuzhou Institute of Technology College of Computing and Information Science Fuzhou350506 China Fuzhou University College of Computer and Data Science Fuzhou350108 China

ISBN: (纸本)9798350356670

Artistic portrait usually uses a set of sparse continuous lines to capture a person's unique appearance characteristics based on real people or photos. This paper addresses the issues of strong background noise and the loss of key details of human faces in artistic portraiture, leading to the inability to produce realistic and high quality portraits. A multi-task portrait painting model with hierarchical loss called WNet is proposed based on U2Net used in salient object detection. This model can generate highly abstract portrait painting that do not depend on the background. Compared to existing popular portrait painting generation methods, the images generated by our method have a better visual effect and retain the key details of the facial features better. The experimental results indicate that the proposed WNet is optimal in terms of PSNR, SSIM and MAE performance. Finally, The software communicating robotic arm can quickly spot-paint the fused art portraits. © 2024 IEEE.

关键词： Robotic arms

来源：评论

学校读者我要写书评

暂无评论

Language Identification System: Employing ReLu for India's Regional Languages (ReLu) 2

Language Identification System: Employing ReLu for India's R...

引用

2nd International Conference for Advancement in Technology, ICONAT 2023

作者： Saini, Lekhraj Defence Institute of Advanced Technology School of Computer Science and Data Science Pune India

ISBN: (纸本)9781665475174

In this paper, I provide a model for language identification that makes use of neural networks. The approach is intended to distinguish between Indian regional and dialectal languages. Individuals of Indian descent are the target population for this model. I train the model using a variety of data sources, including corpora of written and spoken material in a variety of languages. In addition, I use a variety of additional data sources. As a result, we can train the model to be more accurate. When I compare the performance of our model to that of a naive Bayes classifier, I find that the results produced by our model are superior to those produced by the naive Bayes classifier. This is because our model considers more information than the naïve Bayes classifier. The ReLu activation function is employed on each neuron throughout our simulation, and the neural network design comprises several layers. Each neuron receives the ReLu activation function. This allows the model to capture the relationships and correlations that exist between the input data, which improves its capacity to recognize the language employed in a specific sentence. Furthermore, our technology can deal with ambiguous cases, such as sentences written in the devnagri script, which is utilized in a number of Indian languages. This is a benefit provided by our software. Using our system has several advantages, and this is only one of them. © 2023 IEEE.

关键词： Chemical activation

来源：评论

学校读者我要写书评

暂无评论

Uniqueness of the Eigenvalue Inverse Problem for a Class of X-type Matrices with Linear Relations

Uniqueness of the Eigenvalue Inverse Problem for a Class of ...

引用

2023 International Conference on Applied Mathematics and Digital Simulation, AMDS 2023

作者： Yu, Tong Li, Zhibin Wang, Lidong School of Statistics and Data Science Zhuhai Institute of Science and Technology Zhuhai China Dalian Shipping College Dalian China

Constraining a matrix by adding additional conditions to determine the matrix if given eigenpairs is called the eigenvalue inverse problem for matrices. The eigenvalue inverse problem of a matrix can be based on a given combination of different eigenvectors and real numbers to inverse a matrix method, and the method of inversion varies for different types of matrices. In this paper, we investigate the eigenvalue inverse problem for a class of X-type matrices characterized by linear relations, Invert the matrix based on its characteristics, determine eigenvalues and eigenvectors, and finally, prove that the solution of the problem exists and is unique, derive a series of expressions and recursive formulas, and we also check the algorithm's accuracy correctness by giving different instances. © Published under licence by IOP Publishing Ltd.

关键词： Inverse problems

来源：评论

学校读者我要写书评

暂无评论

Road Network Representation Learning with the Third Law of Geography 38

Road Network Representation Learning with the Third Law of G...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Zhou, Haicang Huang, Weiming Chen, Yile He, Tiantian Cong, Gao Ong, Yew-Soon College of Computing and Data Science Nanyang Technological University Singapore Institute of High-Performance Computing Agency for Science Technology and Research Singapore Centre for Frontier Al Research Agency for Science Technology and Research Singapore

Road network representation learning aims to learn compressed and effective vectorized representations for road segments that are applicable to numerous tasks. In this paper, we identify the limitations of existing methods, particularly their overemphasis on the distance effect as outlined in the First Law of Geography. In response, we propose to endow road network representation with the principles of the recent Third Law of Geography. To this end, we propose a novel graph contrastive learning framework that employs geographic configuration-aware graph augmentation and spectral negative sampling, ensuring that road segments with similar geographic configurations yield similar representations, and vice versa, aligning with the principles stated in the Third Law. The framework further fuses the Third Law with the First Law through a dual contrastive learning objective to effectively balance the implications of both laws. We evaluate our framework on two real-world datasets across three downstream tasks. The results show that the integration of the Third Law significantly improves the performance of road segment representations in downstream tasks. Our code is available at https://***/Haicang/Garner. © 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

UniOfGalway@IberLEF 2024:Hope Speech Recognition in Spanish: A Comparative Analysis of Transformer-Based Models 6

UniOfGalway@IberLEF 2024:Hope Speech Recognition in Spanish:...

引用

6th Iberian Languages Evaluation Forum, IberLEF 2024

作者： Subburaj, Arunraj Kathiresan, Amirthagadeshwaran Ponnusamy, Rahul Butilaar, Paul Chakravarthi, Bharathi Raja University of Galway Ireland Data Science Institute University of Galway Ireland School of Computer Science University of Galway Ireland

Hope speech detection in social media is crucial for fostering positive engagement and resilience within marginalized communities. By utilizing machine learning and NLP techniques, researchers have focused on identifying hope speech in online content, achieving F1-score between 76% and 84%. This detection of hope speech contributes to understanding how social media platforms can either create a hopeful or hostile environment, influencing community sentiments and individual identities. Recognizing and promoting hope speech is essential for enhancing interaction quality and support among users, especially within Equality, Diversity, and Inclusion initiatives, emphasizing the significance of hope in digital communication for building inclusive and supportive online ecosystems. © 2024 Copyright for this paper by its authors.

关键词： Contrastive Learning

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：