Recently, several privacy-preserving algorithms for NLP have emerged. These algorithms can be suitable for LLMs as they can protect both training and query data. However, there is no benchmark exists to guide the eval...
详细信息
ISBN:
(纸本)9798350367102;9798350367096
Recently, several privacy-preserving algorithms for NLP have emerged. These algorithms can be suitable for LLMs as they can protect both training and query data. However, there is no benchmark exists to guide the evaluation of these algorithms when applied to LLMs. This paper presents a benchmark framework for evaluating the effectiveness of privacy-preserving algorithms applied to training and query data for fine-tuning LLMs under various scenarios. The proposed benchmark is designed to be transferable, enabling researchers to assess other privacy-preserving algorithms and LLMs. The benchmark focuses on assessing the privacy-preserving algorithms on training and query data when fine-tuning LLMs in various scenarios. We evaluated the SANTEXT+ algorithm on the open-source Llama2-7b LLM using a sensitive medical transcription dataset. Results demonstrate the algorithm's effectiveness while highlighting the importance of considering specific situations when determining algorithm parameters. This work aims to facilitate the development and evaluation of effective privacy-preserving algorithms for LLMs, contributing to the creation of trusted LLMs that mitigate concerns regarding the misuse of sensitive information.
The power system landscape has evolved from isolated end-users to interactive communities due to advances in information and communication technologies. This paper explores peer-to-peer energy (P2PE) trading and shari...
详细信息
The power system landscape has evolved from isolated end-users to interactive communities due to advances in information and communication technologies. This paper explores peer-to-peer energy (P2PE) trading and sharing within a community, where customer incentives for energy exchange enhance collective profits. A two-stage optimization (TSO) framework is proposed: the first stage determines customer participation in P2PE, balancing individual and collective benefits, while the second stage optimizes economic aspects of P2P trading using a payment bargaining model. A case study demonstrates significant cost reductions and improved renewable energy utilization, with notable profit increments for participants. The study highlights the effectiveness of Nash bargaining theory and privacy-preserving algorithms in optimizing social welfare and economic interactions. Limitations include a focus on wind energy and simplified assumptions about energy storage. Future research should incorporate diverse renewable sources, dynamic modeling, and multi-community interactions.
The pervasive issue of fraudulent transactions presents a considerable challenge for financial institutions globally. Developing innovative fraud detection systems is critical to maintaining customer confidence. Howev...
详细信息
ISBN:
(纸本)9783031752001;9783031752018
The pervasive issue of fraudulent transactions presents a considerable challenge for financial institutions globally. Developing innovative fraud detection systems is critical to maintaining customer confidence. However, several factors complicate the creating of effective and efficient fraud detection systems. Notably, fraudulent transactions are infrequent, resulting in imbalanced transaction datasets where legitimate transactions vastly outnumber instances of fraud. This data imbalance can concede the performance of fraud detection. Additionally, stringent data privacy regulations prevent the sharing of customer data, hindering the development of high-performing centralized models. Furthermore, fraud detection mechanisms must remain transparent to avoid impairing the user experience. This research proposes an approach utilizing Federated Learning (FL) with Explainable Artificial Intelligence (XAI) to overcome these obstacles. FL allows financial organizations to train fraud detection models collaboratively without requiring direct data sharing. So, customer confidentiality and data privacy are never compromised. Simultaneously, the incorporation of XAI guarantees that the model's predictions are interpretable by human experts. Experimental evaluations using real-time transaction datasets consistently demonstrate that the FL-based fraud detection system performs well. This study establishes the potential of FL as a reliable, privacy-preserving tool in combating fraud.
The rapid evolution of the Internet of Things (IoT) has revolutionized various sectors, fostering seamless intercommunication and real-time monitoring. Central to this transformation is integrating blockchain technolo...
详细信息
The rapid evolution of the Internet of Things (IoT) has revolutionized various sectors, fostering seamless intercommunication and real-time monitoring. Central to this transformation is integrating blockchain technology, which ensures data integrity and security in IoT networks. This paper provides a meticulous exploration of data aggregation techniques within the context of blockchain-based IoT systems. The study categorizes data aggregation algorithms into privacy-preserving, Machine Learning-Based, Hierarchical, Real-Time, and Custom Aggregation algorithms, each tailored to specific IoT requirements. privacy-preserving Aggregation algorithms focus on safeguarding sensitive data through encryption and secure protocols. Machine Learning-Based Aggregation adapts dynamically to data patterns, offering predictive insights and real-time adaptability. Hierarchical Aggregation organizes devices into a structured hierarchy, optimizing data processing. Real-Time Aggregation processes data instantly, ensuring low latency for time-sensitive applications. Custom Aggregation algorithms are bespoke solutions tailored to unique application demands, emphasizing efficiency and security. Through a comparative analysis of these techniques, this paper explores their advantages, disadvantages, and applicability, addressing the challenges and suggesting future research directions. The integration of blockchain-based data aggregation techniques not only enhances IoT network efficiency but also ensures the longevity and security of modern technological infrastructures. This study builds upon prior research in the field of IoT and blockchain technology by extending the exploration of data aggregation techniques and their implications for network efficiency and security. SLR method has been used to investigate each one in terms of influential properties such as the main idea, advantages, disadvantages, and strategies. The results indicate most of the articles were published in 2021 and 202
Differential privacy is commonly used in the computer science literature as a mathematical definition of privacy for the purpose of quantifying and bounding privacy loss. It induces a preference order over the set of ...
详细信息
Differential privacy is commonly used in the computer science literature as a mathematical definition of privacy for the purpose of quantifying and bounding privacy loss. It induces a preference order over the set of privacy-jeopardizing mechanisms which, in turn, adhere to some properties of this order. We show that a set of five such properties uniquely captures the ordinal implications of prioritizing the alternatives in agreement with differential privacy. The model can also be applied to evaluate the appropriateness of differential privacy in different settings. (C) 2020 Elsevier B.V. All rights reserved.
Mobile Edge Computing (MEC) and Federated Learning (FL) have recently attracted considerable interest for their potential applications across diverse domains. MEC is an architecture for distributed computing that util...
详细信息
Mobile Edge Computing (MEC) and Federated Learning (FL) have recently attracted considerable interest for their potential applications across diverse domains. MEC is an architecture for distributed computing that utilizes computational capabilities near the network edge, enabling quicker data processing and minimizing latency. In contrast, FL is a method in the field of Machine learning (ML) that allows for the simultaneous involvement of multiple participants to collectively train models without revealing their raw data, effectively tackling concerns related to security and privacy. This systematic review explores the core principles, architectures, and applications of FL within MEC and vice versa, providing a comprehensive analysis of these technologies. The study emphasizes FL and MEC's unique characteristics, advantages, and drawbacks, highlighting their attributes and limitations. The study explores the complex architectures of both technologies, showcasing the cutting-edge methods and tools employed for their implementation. Aside from examining the foundational principles, the review explores the depths of the internal mechanisms of FL and MEC, offering a valuable in-depth of their architecture understanding and the fundamental principles and processes that facilitate their operation. At last, the concluding remarks and future research directions are provided.
暂无评论