检索结果-内蒙古大学图书馆

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Hu, Xinyi Lee, Jasper C.H. Lee, Jimmy H.M. Stuckey, Peter J. Department of Computer Science and Engineering The Chinese University of Hong Kong Shatin N.T. Hong Kong Department of Computer Science University of California Davis United States Department of Data Science and AI Monash University Clayton Australia

The recently-proposed framework of Predict+Optimize tackles optimization problems with parameters that are unknown at solving time, in a supervised learning setting. Prior frameworks consider only the scenario where all unknown parameters are (eventually) revealed at the same time. In this work, we propose Multi-Stage Predict+Optimize, a novel extension catering to applications where unknown parameters are instead revealed in sequential stages, with optimization decisions made in between. We further develop three training algorithms for neural networks (NNs) for our framework as proof of concept, all of which can handle mixed integer linear programs. The first baseline algorithm is a natural extension of prior work, training a single NN which makes a single prediction of unknown parameters. The second and third algorithms instead leverage the possibility of updating parameter predictions between stages, and trains one NN per stage. To handle the interdependency between the NNs, we adopt a sequential and parallelized versions of coordinate descent for training. Experimentation on three benchmarks demonstrates the superior learning performance of our methods over classical approaches. © 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

Exploring Deep Learning architectures for Crop and Weed Identification 3

Exploring Deep Learning architectures for Crop and Weed Iden...

引用

3rd International Conference on Artificial Intelligence For Internet of Things, AIIoT 2024

作者： Akaash Dravid, J. Karthik, R.K. Sreekumar, K. School of Computing Amrita Vishwavidyapeetham Department of Computer Science and IT Kochi India

ISBN: (纸本)9798350372120

In agriculture, weeds always prove to be a major threat. It is tedious to do weeding at a later stage when the crops and the weeds are significantly grown. Incorporating technology in agriculture has revolutionized both industries in many ways. In the AI approach towards agriculture, neural networks are employed for identifying and classifying crops and weeds. However, what features they learn, how they learn the difference between crops and weeds is what we need to investigate. Many vegetable crop seedlings and weeds look similar, so a significantly large amount of data is necessary for the model to classify them. In this paper we are using CNN (Convolutional Neural Networks) and Efficientnet architectures to classify the crops and weeds. We further analyze the computational efficiency and model complexity of the architectures, providing insights into their practical applicability in real-world agricultural settings. This helps in better understanding of deep learning techniques and selecting features to perform this task. We have obtained an accuracy of 91% for CNN architecture and 99% for both B3, B4 of EfficientNet Architecture. © 2024 IEEE.

关键词： Crops

来源：评论

学校读者我要写书评

暂无评论

Deep Reinforcement Learning in Stock Trading: Evaluating DDPG and DQN Strategies 6

Deep Reinforcement Learning in Stock Trading: Evaluating DDP...

引用

6th IEEE International Conference on Emerging Smart Computing and Informatics, ESCI 2024

作者： Kodurupaka, Nithin Basavadeepthi, H.M. Pecheti, Shiva Teja Amudha, J. Amrita School of Computing Department of Computer Science and Engineering Bengaluru India

ISBN: (纸本)9798350306613

This study presents a comparative analysis of the Deep Q-Network (DQN) and Deep Deterministic Policy Gradient (DDPG) reinforcement learning algorithms in the context of stock trading, focusing on historical stock price data of S&P 500 stock index. The primary objective is to evaluate the effectiveness of these RL algorithms in making trading decisions and achieving financial performance relative to a passive buy-And-hold strategy. The assessment revolves around key metrics, with a special emphasis on the buy, hold, and sell indications as essential signals for trading actions. Results indicate that DQN outperforms DDPG, showcasing a smaller total loss during the evaluation period. This suggests that DQN may offer superior capabilities in limiting financial losses and making effective trading decisions compared to DDPG, providing valuable insights into algorithm selection for dynamic trading environments. It is crucial to recognize the significance of buy, hold, and sell indications as key elements in assessing the performance of RL algorithms in stock trading scenarios. © 2024 IEEE.

关键词： Algorithmic Trading Buy-And-Hold Strategy DDPG Trading Agent Deep Neural Networks Deep Reinforcement Learning DQN Portfolio Performance Total Return

来源：评论

学校读者我要写书评

暂无评论

Hidden Emotion Detection using Speech, Text, Facial Expressions and Neural Networks 5

Hidden Emotion Detection using Speech, Text, Facial Expressi...

引用

5th IEEE International Conference on Communication, Computing and Industry 6.0, C2I6 2024

作者： Reddy, Ravula Tarun Palaniswamy, Suja Amrita School of Computing Department of Computer Science and Engineering Bengaluru India

ISBN: (纸本)9798350379884

Understanding and recognition of human emotions are very crucial in various fields. This paper proposes a new approach to show the different feelings that are hidden using multi-modalities like video, audio, and textual data. A custom dataset AED-FST has been used in this work which has nearly 317 samples each consisting of some audio clip, some visual media, and a question-and-answer pair. From all these entirely diverse sources, the goal is to create a model that could possibly detect emotional cues. It found hidden emotions with a 67% accuracy, representing how well it could pick up on these weak cues from all different modalities. This method contributes immensely to emotion detection in extremely wide fields, including psychological diagnoses and human-computer interaction, to mention just a few. © 2024 IEEE.

关键词： Human computer interaction

来源：评论

学校读者我要写书评

暂无评论

Household Waste Classification Algorithms Based on Various Convolutional Neural Networks: A Review of Performance Optimization 21

Household Waste Classification Algorithms Based on Various C...

引用

21st International computer Conference on Wavelet Active Media Technology and Information Processing, ICCWAMTIP 2024

作者： Renxiang, Huang Jia, Chen Jianping, Li School of Artificial Intelligence and Big Data Sichuan University of Arts and Science Chengdu635002 China School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu611731 China

ISBN: (纸本)9798331519254

This paper reviews the research progress of deep learning-based household waste classification algorithms. It first introduces the importance of household waste classification and the application value of deep learning in this field. Next, it outlines the theoretical foundations of deep learning technology in waste classification, including the principles of convolutional neural networks and their applications in image recognition. The article provides a detailed analysis of the application of different deep learning models in waste classification, such as MobileNetV2, ShuffleNetV1, EfficientNetV1, and YOLOV5, as well as their optimizations and innovations in practical applications. Furthermore, it discusses optimization and innovation methods for deep learning models, including learning rate adjustment, optimizer selection, attention mechanisms, data augmentation, and lightweight network design. Finally, it summarizes the current research shortcomings and challenges, and looks forward to future research directions. This review provides a systematic perspective for a deeper understanding of the application of deep learning in waste classification and offers directions for future research. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Latent Logic Tree Extraction for Event Sequence Explanation from LLMs 41

Latent Logic Tree Extraction for Event Sequence Explanation ...

引用

41st International Conference on Machine Learning, ICML 2024

作者： Song, Zitao Yang, Chao Wang, Chaojie An, Bo Li, Shuang School of Computer Science and Engineering Nanyang Technological University Singapore School of Data Science The Chinese University of Hong Kong Shenzhen China Skywork AI Singapore

Modern high-stakes systems, such as healthcare or robotics, often generate vast streaming event sequences. Our goal is to design an efficient, plug- and-play tool to elicit logic tree-based explanations from Large Language Models (LLMs) to provide customized insights into each observed event sequence. Built on the temporal point process model for events, our method employs the likelihood function as a score to evaluate generated logic trees. We propose an amortized Expectation-Maximization (EM) learning framework and treat the logic tree as latent variables. In the E-step, we evaluate the posterior distribution over the latent logic trees using an LLM prior and the likelihood of the observed event sequences. LLM provides a high-quality prior for the latent logic trees, however, since the posterior is built over a discrete combinatorial space, we cannot get the closed-form solution. We propose to generate logic tree samples from the posterior using a learnable GFlowNet, which is a diversity-seeking generator for structured discrete variables. The M-step employs the generated logic rules to approximate marginalization over the posterior, facilitating the learning of model parameters and refining the tunable LLM prior parameters. In the online setting, our locally built, lightweight model will iteratively extract the most relevant rules from LLMs for each sequence using only a few iterations. Empirical demonstrations showcase the promising performance and adaptability of our framework. Copyright 2024 by the author(s)

关键词： Trees (mathematics)

来源：评论

学校读者我要写书评

暂无评论

RBLA: Rank-Based-LoRA-Aggregation for Fine-Tuning Heterogeneous Models in FLaaS 31st

RBLA: Rank-Based-LoRA-Aggregation for Fine-Tuning Heterogen...

引用

31st International Conference on Web Services, ICWS 2024, Held as Part of the Services Conference Federation, SCF 2024

作者： Chen, Shuaijun Tavallaie, Omid Nazemi, Niousha Zomaya, Albert Y. School of Computer Science The University of Sydney Camperdown Australia Department of Engineering Science University of Oxford Oxford United Kingdom

ISBN: (纸本)9783031770715

Federated Learning (FL) is a promising privacy-aware distributed learning framework that can be deployed on various devices, such as mobile phones, desktops, and devices equipped with CPUs or GPUs. In the context of server-based Federated Learning as a Service (FLaaS), FL enables a central server to coordinate the training process across multiple devices without direct access to local data, thereby enhancing privacy and data security. Low-Rank Adaptation (LoRA) is a method that efficiently fine-tunes models by focusing on a low-dimensional subspace of the model’s parameters. This approach significantly reduces computational and memory costs compared to fine-tuning all parameters from scratch. When integrated with FL, particularly in a FLaaS environment, LoRA allows for flexible and efficient deployment across diverse hardware with varying computational capabilities by adjusting the local model’s rank. However, in LoRA-enabled FL, different clients may train models with varying ranks, which poses challenges for model aggregation on the server. Current methods for aggregating models of different ranks involve padding weights to a uniform shape, which can degrade the global model’s performance. To address this issue, we propose Rank-Based LoRA Aggregation (RBLA), a novel model aggregation method designed for heterogeneous LoRA structures. RBLA preserves key features across models with different ranks. This paper analyzes the issues with current padding methods used to reshape models for aggregation in a FLaaS environment. Then, we introduce RBLA, a rank-based aggregation method that maintains both low-rank and high-rank features. Finally, we demonstrate the effectiveness of RBLA through comparative experiments with state-of-the-art methods. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

An optimal hybrid cascade regional convolutional network for cyberattack detection

An optimal hybrid cascade regional convolutional network for...

引用

作者： Alqahtani, Ali Khan, Surbhi Bhatia Department of Networks and Communications Engineering College of Computer Science and Information Systems Najran University Najran Saudi Arabia Department of Data Science School of Science Engineering and Environment University of Salford Salford United Kingdom Department of Electrical and Computer Engineering Lebanese American University Byblos Lebanon

Cyber-physical systems (CPS) and the Internet of Things (IoT) technologies link urban systems through networks and improve the delivery of quality services to residents. To enhance municipality services, information and communication technologies (ICTs) are integrated with urban systems. However, the large number of sensors in a smart city generates a significant amount of delicate data, like medical records, credit card numerics, and location coordinates, which are transported across a network to data centers for analysis and processing. This makes smart cities vulnerable to cyberattacks because of the resource constraints of their technology infrastructure. Applications for smart cities pose many security challenges, such as zero-day attacks resulting from exploiting weaknesses in various protocols. Therefore, this paper proposes an optimal hybrid transit search-cascade regional convolutional neural network (hybrid TS-Cascade R-CNN) to detect cyberattacks. The proposed model combines the hybrid transit-search approach with the cascade regional convolutional neural network to create an optimal solution for cyberattack detection. The cascade regional convolutional network uses a hybrid transit search algorithm to enhance the effectiveness of cyberattack detection. By integrating these two approaches, the system can leverage both global traffic patterns and local indicators to improve the accuracy of attack detection. During the training process, the proposed model recognizes and classifies malicious input even in the presence of sophisticated attacks. Finally, the experimental analysis is carried out for various attacks based on different metrics. The accuracy rate attained by the proposed approach is 99.2%, which is acceptable according to standards. © 2023 John Wiley & Sons Ltd.

关键词： Zero-day attack

来源：评论

学校读者我要写书评

暂无评论

Hybrid Water Quality Prediction With Multimodal Low-Rank Fusion and Localized Attention

引用

IEEE Internet of Things Journal 2025年第12期12卷 21158-21169页

作者： Bi, Jing Li, Yibo Yuan, Haitao Wang, Mengyuan Wang, Ziqi Zhang, Jia Zhou, Mengchu Beijing University of Technology College of Computer Science Beijing Laboratory of Smart Environmental Protection Beijing100124 China Beihang University School of Automation Science and Electrical Engineering Beijing100191 China Beihang University School of Energy and Power Engineering Beijing100191 China Southern Methodist University Department of Computer Science DallasTX75206 United States New Jersey Institute of Technology Department of Electrical and Computer Engineering NewarkNJ07102 United States

Water quality prediction methods forecast the short- or long-term trends of its changes, providing proactive advice for preventing and controlling water pollution. Existing water quality prediction methods typically fail to capture water quality’s nonlinear characteristics accurately and only consider historical time series data. However, meteorology and other factors also significantly impact water quality indicators. Therefore, considering only historical data of water quality time series is not feasible. To solve this problem, this work proposes a hybrid water quality prediction model called CMLIP, which integrates convNeXt V2, multimodal bottleneck transformer, low-rank multimodal fusion, iTransformer, and PatchTST. CMLIP inputs water quality time series and meteorological remotely sensed rainfall images into a multimodal fusion module before prediction. Specifically, CMLIP integrates the model of ConvNeXt V2 to extract image features. Its multimodal fusion module combines a multimodal bottleneck transformer and the low-rank multimodal fusion to fuse the time series and images. Furthermore, CMLIP combines iTransformer and PatchTST to form an improved prediction module that realizes the prediction of fused features. Experimental results with real-life water quality time series and remotely sensed rainfall images demonstrate that CMLIP when fusing meteorological data, achieves an average improvement of 17% in water quality forecasting accuracy compared to forecasts using only water quality time series. Moreover, CMLIP outperforms other state-of-the-art algorithms in both data fusion and prediction, with an average enhancement of 6% in fusion effectiveness and an average improvement of 22% in prediction accuracy. © 2014 IEEE.

关键词： Rain

来源：评论

学校读者我要写书评

暂无评论

SELF-Clustering Graph Transformer Approach to Model Resting State Functional Brain Activity 22

SELF-Clustering Graph Transformer Approach to Model Resting ...

引用

22nd IEEE International Symposium on Biomedical Imaging, ISBI 2025

作者： Thapaliya, Bishal Akbas, Esra Sapkota, Ram Ray, Bhaskar Calhoun, Vince Liu, Jingyu Georgia State University Department of Computer Science Atlanta United States Tri-Institutional Center for Translational Research in Neuroimaging and Data Science United States School of Electrical and Computer Engineering Georgia Institute of Technology Atlanta United States

ISBN: (纸本)9798331520526

Resting-state functional magnetic resonance imaging (rs-fMRI) offers valuable insights into the human brain's functional organization and is a powerful tool for investigating the relationship between brain function and cognitive processes as it allows for the functional organization of the brain to be captured without relying on a specific task or stimuli. In this study, we introduce a novel attention mechanism for graphs with subnetworks, named Self Clustering Graph Transformer (SCGT), designed to handle the issue of uniform node updates in graph transformers. By using static functional connectivity (FC) correlation features as input to the transformer model, SCGT effectively captures the sub-network structure of the brain by performing cluster-specific updates to the nodes unlike uniform node updates like vanilla graph transformers, further allowing us to learn and interpret the sub clusters. We validate our approach on the Adolescent Brain Cognitive Development (ABCD) dataset, comprising 7,957 participants, for the prediction of total cognitive score and gender classification. Our results demonstrate that SCGT outperforms the vanilla graph transformer method, and other recent models, offering a promising tool for modeling brain functional connectivity and interpreting the underlying subnetwork structures. © 2025 IEEE.

关键词： Brain mapping

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：