Abstract:The conventional machine learning process typically operates under the premise of centralized data aggregation, where all data is collected at a central location for model training. However, this raises substantial privacy concerns when the data contains private information. In this context, federated learning has emerged as a prominent solution for preserving privacy in machine learning. This innovative paradigm allows multiple data owners, or clients, to collaboratively train a machine learning model while keeping their local data unshared. A federated learning task is typically initiated by companies, often referred to as model owners, who do not possess enough training data and are willing to financially remunerate clients who contribute to the development of the federated learning model. This situation demands a trading platform that enables model owners to effectively select clients, while ensuring robustness against malicious clients who execute poisoning attacks for unfair financial gain. To address these issues, we design a contribution-based exploration-exploitation mechanism implemented as a smart contract. This mechanism cherry-picks clients with high data quality based on the Shapley value, which is calculated based on local models to evaluate the contribution of each client. Unlike other state-of-the-art security mechanisms, the proposed mechanism can adapt to various scenarios with heterogeneous data distribution and various attacks, while mitigating the effect of malicious behaviors without compromising training accuracy. To accelerate the time-consuming calculation of the Shapley value, we design a parallel computing algorithm that partitions blockchain nodes into multiple shards and distributes calculation tasks among them. The algorithm improves efficiency and tolerates potential false calculation results from malicious nodes.

Decentralized and Secure Deduplication with Dynamic Ownership in MLaaS

Enabling Secure Deduplication in Encrypted Decentralized Storage

Updatable Block-Level Deduplication with Dynamic Ownership Management on Encrypted Data

Two-party Interactive Secure Deduplication with Efficient Data Ownership Management in Cloud Storage

A Secure and Efficient Data Deduplication Scheme with Dynamic Ownership Management in Cloud Computing

CSED: Client-Side Encrypted Deduplication Scheme Based on Proofs of Ownership for Cloud Storage

Secure and Efficient Cloud Data Deduplication with Ownership Management

Anonymous deduplication of encrypted data with proof of ownership in cloud storage

Privacy-preserving Machine Learning as a Service

Distributed Machine Learning Oriented Data Integrity Verification Scheme in Cloud Computing Environment

DedupDUM: Secure and Scalable Data Deduplication with Dynamic User Management

Private, Yet Practical, Multiparty Deep Learning.

Blockchain-Based platform for Distribution AI

Binary Linear Tree Commitment-based Ownership Protection for Distributed Machine Learning

Golden Grain: Building a Secure and Decentralized Model Marketplace for MLaaS

De-duplication of Encrypted Data with Proof of Ownership in Cloud Storage

Secure and Efficient Cloud Data Deduplication with Randomized Tag.

Enabling Transparent Deduplication and Auditing for Encrypted Data in Cloud

Towards blockchain-enabled decentralized and secure federated learning

Privacy-preserving Decentralized Federated Deep Learning.

TDML -- A Trustworthy Distributed Machine Learning Framework