Privacy-Preserving Techniques in Machine Learning and Data Analytics

The field of privacy-preserving techniques in machine learning and data analytics is rapidly advancing, with a focus on developing innovative methods to protect sensitive information while enabling secure and efficient data analysis. Recent developments have seen a surge in the use of homomorphic encryption, secure multi-party computation, and differential privacy to preserve data confidentiality and privacy in various applications, including financial transactions, smart meter data analysis, and language model inference. Noteworthy papers in this area include:

Integrating Building Thermal Flexibility Into Distribution System: A Privacy-Preserved Dispatch Approach, which proposes a novel privacy-preserved optimal dispatch approach for distribution systems incorporating buildings, and
Comet: Accelerating Private Inference for Large Language Model by Predicting Activation Sparsity, which presents an efficient private inference system that exploits the spatial locality of predicted sparse distributions to avoid computations involving zero values.

Sources

Privacy-Preserving Credit Card Approval Using Homomorphic SVM: Toward Secure Inference in FinTech Applications

Integrating Building Thermal Flexibility Into Distribution System: A Privacy-Preserved Dispatch Approach

Cryptanalysis of a Lattice-Based PIR Scheme for Arbitrary Database Sizes

Enhancing Noisy Functional Encryption for Privacy-Preserving Machine Learning

DPolicy: Managing Privacy Risks Across Multiple Releases with Differential Privacy

Comet: Accelerating Private Inference for Large Language Model by Predicting Activation Sparsity

Private LoRA Fine-tuning of Open-Source LLMs with Homomorphic Encryption

Revenue Optimization in Video Caching Networks with Privacy-Preserving Demand Predictions

Privacy-Preserving Analytics for Smart Meter (AMI) Data: A Hybrid Approach to Comply with CPUC Privacy Regulations

Multiparty Selective Disclosure using Attribute-Based Encryption

Private Transformer Inference in MLaaS: A Survey