Distributed cluster systems have emerged as a useful way to perform large scale computations. These clusters can be rented at service providers, which imposes constraints regarding the privacy of the data the computat...
详细信息
Distributed cluster systems have emerged as a useful way to perform large scale computations. These clusters can be rented at service providers, which imposes constraints regarding the privacy of the data the computations are done on. Fur- thermore, straggling servers that take an excessive amount of time to? nish their computation slow down the overall process gravely. In this thesis, a coding scheme that mitigates the e? ect of straggling servers and assures privacy is introduced. This is made possible by a concatenation of two Reed-Solomon codes, one for straggler mitigation directly at the service providers and one in combination with random data to assure privacy which is used on the data itself, before it is distributed to multiple service providers. The impact of di? erent network structures on the overall runtime as well as the impact of privacy on the optimal code rate for minimizing the overall runtime is investigated and set in relation with existing theory, such as studies from Lee et al. on the optimal code rate of maximum distance separable codes for distributed computing.
Fueled by its successful commercialization, the recommender system (RS) has gained widespread attention. However, as the training data fed into the RS models are often highly sensitive, it ultimately leads to severe p...
详细信息
Fueled by its successful commercialization, the recommender system (RS) has gained widespread attention. However, as the training data fed into the RS models are often highly sensitive, it ultimately leads to severe privacy concerns, especially when data are shared among different platforms. In this paper, we follow the tune of existing works to investigate the problem of secure sparse matrix multiplication for cross-platform RSs. Two fundamental and critical issues are addressed: preserving the training data privacy and breaking the data silo problem. Specifically, we propose two concrete constructions with significantly boosted efficiency. They are designed for the sparse location insensitive case and location sensitive case, respectively. State-of-the-art cryptography building blocks including homomorphic encryption (HE) and private information retrieval (PIR) are fused into our protocols with non-trivial optimizations. As a result, our schemes can enjoy the HE acceleration technique without privacy trade-offs. We give formal security proofs for the proposed schemes and conduct extensive experiments on both real and large-scale simulated datasets. Compared with state-of-the-art works, our two schemes compress the running time roughly by $10\times$10x and $2.8\times$2.8x. They also attain up to $15\times$15x and $2.3\times$2.3x communication reduction without accuracy loss.
The Industrial Internet of Things (IIoT) leads to increasingly-interconnected industrial processes and environments, which, in turn, result in stakeholders collecting a plethora of information. Even though the global ...
详细信息
ISBN:
(纸本)9798350327939;9798350327946
The Industrial Internet of Things (IIoT) leads to increasingly-interconnected industrial processes and environments, which, in turn, result in stakeholders collecting a plethora of information. Even though the global sharing of information and industrial collaborations in the IIoT promise significant improvements concerning productivity, sustainability, and product quality, among others, the majority of stakeholders is hesitant to implement them due to confidentiality and reliability concerns. However, strong technical guarantees could convince them of the contrary. Thus, to address these concerns, our interdisciplinary efforts focus on establishing and realizing secure industrial collaborations in the IIoT. By applying private computing, we are indeed able to reliably secure collaborations that not only scale to industry-sized applications but also allow for use case-specific confidentiality guarantees. Hence, improvements that follow from industrial collaborations with (strong) technical guarantees are within reach, even when dealing with cautious stakeholders. Still, until we can fully exploit these benefits, several challenges remain, primarily regarding collaboration management, introduced overhead, interoperability, and universality of proposed protocols.
This paper meticulously examines the privacy challenges stemming from the escalating volume of data within the new power systems. It thoroughly explores various privacy computing techniques, including encryption and s...
详细信息
ISBN:
(纸本)9789819743896;9789819743902
This paper meticulously examines the privacy challenges stemming from the escalating volume of data within the new power systems. It thoroughly explores various privacy computing techniques, including encryption and statistical security, with a particular emphasis on applications such as homomorphic encryption, federated learning and differential privacy within the power grid. The analysis underscores the pivotal role of privacy computing in addressing privacy-related concerns within the power industry, particularly focusing on issues like data transmission security and multi-party collaboration. Furthermore, the paper proposes innovative solutions aimed at bolstering privacy protection, fostering data sharing initiatives, and enhancing cost-effectiveness through the strategic implementation of privacy computing technologies. Recognizing the paramount importance of understanding privacy protection in the context of the digital transformation of power systems, this paper seeks to provide valuable insights to guide further research and application of privacy computing. Ultimately, these insights aim to facilitate secure data circulation within the energy sector.
Batch schemes provide a way to amortize the cost of NP verifiers across multiple instances. In this work, we introduce an instance batch scheme that allows a prover to iteratively prove the correctness of multiple exe...
详细信息
ISBN:
(纸本)9789819751006;9789819751013
Batch schemes provide a way to amortize the cost of NP verifiers across multiple instances. In this work, we introduce an instance batch scheme that allows a prover to iteratively prove the correctness of multiple executions of a function F. For N instance-witness tuples where each represents one correct execution of F on given inputs, our scheme could batch them into two tuples and therefore reduce the task of verifying these tuples into the task of verifying only two tuples. Specifically, the prover complexity is O(N) multi-exponentiations of size vertical bar F vertical bar, the verifier complexity is O(vertical bar F vertical bar) field multiplications, and the proof size is O(vertical bar F vertical bar) field elements, where vertical bar F vertical bar| denotes the size of F. Moreover, we provide a technique that allows multiple provers to generate a proof parallelly, which would accelerate the proof generation process in practice. We apply our batch scheme to the Decentralized private Computation (DPC) scenario and implement this application. The benchmark results show that the proving time has been reduced by approximately 9% similar to 11% compared with the state-of-the-art DPC scheme.
Nowadays, with tremendous visual media stored and even processed in the cloud, the privacy of visual media is also exposed to the cloud. In this paper we propose a private face identification method based on sparse re...
详细信息
ISBN:
(纸本)9783319254173;9783319254166
Nowadays, with tremendous visual media stored and even processed in the cloud, the privacy of visual media is also exposed to the cloud. In this paper we propose a private face identification method based on sparse representation. The identification is done in a secure way which protects both the privacy of the subjects and the confidentiality of the database. The face identification server in the cloud contains a list of registered faces. The surveillance client captures a face image and require the server to identify if the client face matches one of the suspects, but otherwise reveals no information to neither of the two parties. This is the first work that introduces sparse representation to the secure protocol of private face identification, which reduces the dimension of the face representation vector and avoid the patch based attack of a previous work. Besides, we introduce a secure Euclidean distance algorithm for the secure protocol. The experimental results reveal that the cloud server can return the identification results to the surveillance client without knowing anything about the client face image.
private computing on Public Platforms (PCPP) is a new technology designed to enable secure and private execution of applications on remote, potentially hostile, public platforms. PCPP uses a host assessment to validat...
详细信息
ISBN:
(纸本)9781424423248
private computing on Public Platforms (PCPP) is a new technology designed to enable secure and private execution of applications on remote, potentially hostile, public platforms. PCPP uses a host assessment to validate a host's hardware and software configuration and then uses four active security building blocks which together allow an application to remain unaltered, unmonitored, and unrecorded before, during, and after execution on the public platform. In this paper we develop a key PCPP building block, Secure Context Switch (SCS), which isolates an executing application's context, i.e. its executable code, data segments, heap, and stack, using encryption techniques. Additionally, we detail our implementation of SCS and offer experimental results showing the performance impact of protecting an application with SCS.
暂无评论