IEEE Intelligent Systems · 2020 · 612 citations · 21 references
ComputingComputer ScienceInformation TechnologyComputer SecurityMachine LearningStatistical LearningData PrivacyData ProtectionFederated LearningSecure Data SharingSecure Multi-party ComputationTransfer LearningFederated Transfer LearningLabelsPrivacy-preserving LearningPrivacy-preserving Machine LearningPrivacy-preserving ProcessingSecure Data FederationSecure Data TransferSecure Machine Learning
Machine learning requires large amounts of data, yet data are often scattered across organizations and cannot be easily integrated due to legal and practical constraints. The paper introduces federated transfer learning (FTL) to improve statistical models in a federated data setting. FTL shares knowledge across parties without compromising privacy, transfers complementary knowledge, and incorporates a secure transfer cross‑validation approach to protect performance. The framework enables target parties to build more flexible and powerful models using rich source labels, achieves comparable accuracy to non‑privacy‑preserving methods with minimal model changes, and is adaptable to various secure multi‑party machine learning tasks.
Machine learning relies on the availability of a vast amount of data for training. However, in reality, most data are scattered across different organizations and cannot be easily integrated under many legal and practical constraints. In this paper, we introduce a new technique and framework, known as federated transfer learning (FTL), to improve statistical models under a data federation. The federation allows knowledge to be shared without compromising user privacy, and enables complimentary knowledge to be transferred in the network. As a result, a target-domain party can build more flexible and powerful models by leveraging rich labels from a source-domain party. A secure transfer cross validation approach is also proposed to guard the FTL performance under the federation. The framework requires minimal modifications to the existing model structure and provides the same level of accuracy as the non-privacy-preserving approach. This framework is very flexible and can be effectively adapted to various secure multi-party machine learning tasks.
21
ImageNet: A large-scale hierarchical image database
Jia Deng, Wei Dong, Richard Socher et al. · 2009 IEEE Conference on Computer Vision and Pattern Recognition · 2009 · 61.8K citations
Adi Shamir · Communications of the ACM · 1979 · 13.5K citations · Full text
Practical Secure Aggregation for Privacy-Preserving Machine Learning
Keith Bonawitz, Vladimir Ivanov, Ben Kreuter et al. · 2017 · 3.4K citations
Tat‐Seng Chua, Jinhui Tang, Richang Hong et al. · 2009 · 3K citations
Privacy-Preserving Deep Learning
Reza Shokri, Vitaly Shmatikov · 2015 · 2.3K citations