SpinML: Customized Synthetic Data Generation for Private Training of Specialized ML Models
This work addresses the challenge of training personalized ML models on smart devices for users who need specialized services but face data scarcity and privacy concerns, representing an incremental improvement in synthetic data generation methods.
The authors tackled the problem of training specialized ML models without sufficient labeled data or access to private user data by proposing SpinML, a system that generates customized synthetic image data from a few sanitized reference images, enhancing model performance while respecting privacy preferences.
Specialized machine learning (ML) models tailored to users needs and requests are increasingly being deployed on smart devices with cameras, to provide personalized intelligent services taking advantage of camera data. However, two primary challenges hinder the training of such models: the lack of publicly available labeled data suitable for specialized tasks and the inaccessibility of labeled private data due to concerns about user privacy. To address these challenges, we propose a novel system SpinML, where the server generates customized Synthetic image data to Privately traIN a specialized ML model tailored to the user request, with the usage of only a few sanitized reference images from the user. SpinML offers users fine-grained, object-level control over the reference images, which allows user to trade between the privacy and utility of the generated synthetic data according to their privacy preferences. Through experiments on three specialized model training tasks, we demonstrate that our proposed system can enhance the performance of specialized models without compromising users privacy preferences.