检索结果-内蒙古大学图书馆

22nd Wind and Solar Integration Workshop, WIW 2023

作者： Forbes, Kevin F. Energy and Environmental Data Science Malahide Ireland

ISBN: (纸本)9781839539824

Using data from Great Britain, this paper begins by observing that the current solar energy forecasts are inaccurate even though the capacity-weighted metric of the forecast error suggests otherwise. This occurs because weighting the mean absolute error by capacity will yield a biased assessment of the forecast error when technologies with a low capacity factor, such as solar energy generation, are evaluated. Consistent with the view that the capacity-weighted metric of the forecast error can lead to a misleading assessment of forecast accuracy, the paper’s analysis indicates that the forecasts do not fully reflect expected weather information. It is also observed that solar energy generation is highly autoregressive, which can be modeled to improve predictability. An Autoregressive Moving Average with Exogenous Inputs (ARMAX) model is formulated based on those properties. The modelwas estimated using 30-minute data from Dec 7, 2018, through Dec 31, 2021. The model is evaluated using out-of-sample data from Jan 1, 2022, to Jun 30, 2022. The period-ahead out-of-sample predictions have a weighted-mean-absolute-percentage-error (WMAPE) of about 3.3 %, substantially less than the approximately 14.7% WMAPE associated with the solar energy forecasts used by the system operator over the same period. © Energynautics GmbH.

关键词： Solar energy

来源：评论

学校读者我要写书评

暂无评论

Peephole Optimization Strategies for Intermediate Code in the Compilation Process 1

Peephole Optimization Strategies for Intermediate Code in th...

引用

1st International Conference on Intelligent Manufacturing and Cloud Computing, ICIMCC 2024

作者： Guan, Yunxin Yun, Jing Sun, Pengfei Su, Xiaoming College of Data Science and Application Inner Mongolia University of Technology Hohhot China

ISBN: (纸本)9781643685779

In the design of modern compilers, the generation and optimization of intermediate code play a crucial role. Serving as a bridge between source code and target machine code, intermediate code provides ample opportunities for optimization. This paper focuses on porthole optimization techniques, proposing a new optimization strategy aimed at reducing unnecessary computations and memory accesses by analyzing control and data flows within intermediate code, thereby enhancing program execution efficiency. We first outline the various stages of the compilation process, followed by a detailed exploration of the basic principles and implementation methods of porthole optimization. Finally, we validate the effectiveness and advantages of this optimization strategy through experiments. The experimental results indicate significant improvements in the execution performance of intermediate code after the adoption of porthole optimization, providing new insights and directions for compiler optimization. © 2025 The Authors.

关键词： Program compilers

来源：评论

学校读者我要写书评

暂无评论

Preserving Privacy in Healthcare Analytics: Federated Learning for Breast Cancer Prediction in a Collaborative Learning Framework 15

Preserving Privacy in Healthcare Analytics: Federated Learni...

引用

15th International Conference on Advances in Computing, Control, and Telecommunication Technologies, ACT 2024

作者： Anuvarshini, G.B. Varshaa, D. Sujithra, M. Devi, D. Sudha Department of Computing- Data Science Coimbatore Institute of Technology Coimbatore India

ISBN: (纸本)9798331300579

Federated learning has emerged as a promising technique in machine learning, enabling collaborative training across distributed datasets. Particularly in fields like healthcare, where data privacy is paramount, federated learning offers a means to improve privacy and security. This paper addresses the critical need for safeguarding sensitive medical data by proposing a privacy-preserving federated learning algorithm. Leveraging homomorphic encryption, the proposed algorithm ensures that the learning process protects both data and model integrity. By employing a secure multi-party computation protocol, the algorithm defends against potential adversaries seeking to compromise the deep learning model. Evaluation on real-world medical datasets demonstrates the effectiveness of the approach in preserving privacy and security amidst collaborative learning environments. © Grenze Scientific Society, 2024.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

The Application of Adversarial Training Based on Gradient Constraint Optimization Method to Sentiment Analysis

Journal of Network Intelligence

引用

Journal of Network Intelligence 2024年第1期9卷 587-598页

作者： Xie, Zhichun Liu, Jianhua Hu, Renyuan Wang, Jiacan Wang, Xiaofeng School of Data and Computer Science Xiamen Institute of Technology Xiamen361021 China School of Computer Science and Mathematics Fujian University of Technology Fuzhou350118 China

The previous adversarial training models failed to pay attention to the influence of the changing gradient of the loss function in the current training on the model. The perturbation injected into the model is only processed by standardization and constraints. This paper proposes a two-way encoding of a converter based on a fast gradient descent of a constraint strategy. The device characterizes the technical model and applies it to sentiment analysis tasks. The model can adapt to gradient changes, and can improve the model’s convergence ability in the early stage of training. In the later stage of training, the perturbation can be appropriately increased according to the gradient to improve the robustness of the model. The experimental results show that even in the expectations of different languages, the classification accuracy of the model is significantly improved compared with the previous transfer learning models and adversarial training. © 2024, Taiwan Ubiquitous Information CO LTD. 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

Privacy-preserving vulnerability detection scheme for smart contracts 3

Privacy-preserving vulnerability detection scheme for smart ...

引用

3rd International Conference on Machine Vision, Automatic Identification, and Detection, MVAID 2024

作者： Shao, Xuping Ou, Zhenyang Faculty of Data Science City University of Macau Macau999078 China

ISBN: (纸本)9781510681880

Smart contracts are automated agreements encoded in the form of code, enabling functionalities such as automated asset transfers and digital asset issuance. These contracts eliminate the need for third-party trust verification, thereby reducing transactional intermediates and costs. Due to their irrevocability and non-breachability, smart contracts execute automatically once preset conditions are met, without the possibility of revocation or modification. However, these characteristics can be exploited by malicious individuals seeking illegitimate gains through vulnerabilities in the contracts. Existing smart contract vulnerability detection tools primarily rely on expert-defined rules, leading to a high false alarm rate and a tedious detection process. Nevertheless, smart contract owners often withhold open-sourcing their contracts due to privacy concerns. This reluctance results in issues like insufficient data and incomplete training samples for deep learning models during the training phase. Consequently, the trained models suffer from limited detection scope and inadequate precision. We compare our approach with existing deep learning methods for detecting vulnerabilities in smart contracts: BiLSTM+Attention. To address these challenges, this paper introduces a federated learning-based approach for detecting vulnerabilities in smart contracts. This method leverages Federated Learning and the BERT model for smart contract vulnerability detection. Significantly, enterprises involved in this approach are not required to provide source code;instead, they participate in federated learning by contributing gradient data. The BERT model is then trained on this gradient dataset. Experimental results demonstrate that the proposed method achieves a remarkable 91.36% accuracy in vulnerability detection. © 2024 SPIE.

关键词： Smart contract

来源：评论

学校读者我要写书评

暂无评论

Prompt Engineering for Curriculum Design 4

Prompt Engineering for Curriculum Design

引用

4th International Conference on Educational Technology, ICET 2024

作者： Leung, Jonathan Shen, Zhiqi Nanyang Technological University College of Computing and Data Science Singapore Singapore

ISBN: (纸本)9798350376944

Large Language Models (LLMs) have recently demonstrated successes in a broad range of areas that require language understanding and generation. The design of course curricula is a key part of the educational process, as it helps students understand expectations and learning goals, and teachers to maintain consistency when conducting the course. Thus, there would be benefits if educators could leverage the capabilities of LLMs. The first benefit is the potential for improving course content, which would lead to better learning outcomes for students. Secondly, educators can save time by having an LLM assist in creating course material or assessments. However, many methods of using LLMs in specific applications fine-tune LLMs, which requires task-specific data as well as technical knowledge to perform the fine-tuning. In this work we use Prompt Engineering, a set of methods that aims to improve generated responses from LLMs by altering the input prompt to the LLM. Often, however, advice for designing prompts is very broad, such as "be concise". We utilize prompt patterns to create and propose three reusable, specific, and customizable prompts that can be used to assist in the design of a course syllabus, lesson material, and assessment questions. The use of prompt patterns additionally aims to produce reliable and consistent results from LLMs. © 2024 IEEE.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Not So Atomic Payments: A Supply Chain Example 20

Not So Atomic Payments: A Supply Chain Example

引用

20th IEEE International Conference on Mobile Ad Hoc and Smart Systems, MASS 2023

作者： De Silva, Anupa Thakur, Subhasis Breslin, John University of Galway Data Science Institute Ireland

ISBN: (纸本)9798350324334

The advent of blockchain as a technology is revolutionizing the supply chain domain with its tamper-proof and traceable architecture. However, the blockchain must be fed with accurate data to actualize its immutability, especially in supply chains where data origin is outside the blockchain. When the data sources (e.g., IoT devices) are managed by the same entity that intends to prove the proposition built on top of those data causes a conflict of interest and undermines the data's credibility. Therefore, the independence of data oracles is crucial, and it allows supply chain stakeholders to focus on their work as well. Using the payment channel network (PCN) concept, we propose a low-cost payment protocol for supply chains with an independent and trusted data oracle environment. It ensures the atomicity of the end-to-end payment from a consumer to the manufacturer, including other participants and data oracles. Further, if the primary transaction fails, the Oracle environment is still compensated regardless, and the consumer is not held liable for that payment. We present how to achieve this unique type of atomicity requirement and discuss its potential, security, and future direction. © 2023 IEEE.

关键词： Supply chains

来源：评论

学校读者我要写书评

暂无评论

Combining Checkpoint/Restart and Replication for Fault Tolerance with High Performance 31

Combining Checkpoint/Restart and Replication for Fault Toler...

引用

31st IEEE International Conference on High Performance Computing, data, and Analytics Workshops, HiPCW 2024

作者： Joshi, Sarthak Vadhiyar, Sathish Department of Computational and Data Sciences Indian Institute of Science Bengaluru India

ISBN: (纸本)9798331509118

As we have entered Exascale computing, the faults in high-performance systems are expected to increase considerably. To compensate for a higher failure rate, the standard checkpoint/restart technique would need to create checkpoints at a much higher frequency, resulting in an excessive amount of overhead, which would not be sustainable for many scientific applications. To improve application efficiency in such high-failure environments, the mechanism of replication of MPI processes was proposed. Replication allows for fast recovery from failures by simply dropping the failed processes and using their replicas to continue the regular operation of the application. We have implemented FTHP-MPI (Fault Tolerance and High-Performance MPI), a novel fault-tolerant MPI library that augments checkpoint/restart with replication to provide resilience from failures. The novelty of our work is that it is designed to provide fault tolerance in a native MPI library. This lets application developers achieve fault tolerance at high failure rates while also using efficient communication protocols in the native MPI libraries that are generally fine-tuned for specific HPC platforms. Our work employs various concepts, including MPI-agnostic checkpointing from DMTCP and communicator shrinking from ULFM, and implements them together without any code modifications on the user end. We have also implemented efficient parallel communication techniques that involve replicas. © 2024 IEEE

关键词： Failure rate

来源：评论

学校读者我要写书评

暂无评论

HyperLogic: Enhancing Diversity and Accuracy in Rule Learning with HyperNets 38

HyperLogic: Enhancing Diversity and Accuracy in Rule Learnin...

引用

38th Conference on Neural Information Processing Systems, NeurIPS 2024

作者： Yang, Yang Ren, Wendi Li, Shuang School of Data Science The Chinese University of Hong Kong Shenzhen China

Exploring the integration of if-then logic rules within neural network architectures presents an intriguing area. This integration seamlessly transforms the rule learning task into neural network training using backpropagation and stochastic gradient descent. From a well-trained sparse and shallow neural network, one can interpret each layer and neuron through the language of logic rules, and a global explanatory rule set can be directly extracted. However, ensuring interpretability may impose constraints on the flexibility, depth, and width of neural networks. In this paper, we propose HyperLogic: a flexible approach leveraging hypernetworks to generate weights of the main network. HyperLogic can be combined with existing differentiable rule learning methods to generate diverse rule sets, each capable of capturing heterogeneous patterns in data. This provides a simple yet effective method to increase model flexibility and preserve interpretability. We theoretically analyze the benefits of the HyperLogic by examining the approximation error and generalization capabilities under two types of regularization terms: sparsity and diversity regularization. Experiments on real data demonstrate that our method can learn more diverse, accurate, and concise rules. Our code is publicly available at https://***/YangYang-624/HyperLogic. © 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

UNCERTAINTY-AWARE CONSTRAINT INFERENCE IN INVERSE CONSTRAINED REINFORCEMENT LEARNING 12

UNCERTAINTY-AWARE CONSTRAINT INFERENCE IN INVERSE CONSTRAINE...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Xu, Sheng Liu, Guiliang School of Data Science The Chinese University of Hong Kong Shenzhen China

Aiming for safe control, Inverse Constrained Reinforcement Learning (ICRL) considers inferring the constraints respected by expert agents from their demonstrations and learning imitation policies that adhere to these constraints. While previous ICRL works often neglected underlying uncertainties during training, we contend that modeling these uncertainties is crucial for facilitating robust constraint inference. This insight leads to the development of an Uncertainty-aware Inverse Constrained Reinforcement Learning (UAICRL) algorithm. Specifically, 1) aleatoric uncertainty arises from the inherent stochasticity of environment dynamics, leading to constraint-violating behaviors in imitation policies. To address this, UAICRL constructs risk-sensitive constraints by incorporating distributional Bellman updates into the cumulative costs model. 2) Epistemic uncertainty, resulting from the model's limited knowledge of Out-of-Distribution (OoD) samples, affects the accuracy of step-wise cost predictions. To tackle this issue, UAICRL develops an information-theoretic quantification of the epistemic uncertainty and mitigates its impact through flow-based generative data augmentation. Empirical results demonstrate that UAICRL consistently outperforms other baselines in continuous and discrete environments with stochastic dynamics. The code is available at https://***/Jasonxu1225/UAICRL. © 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

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：