Auditing Data Provenance in Text-Generation Models
arXiv:1811.00513
Abstract
To help enforce data-protection regulations such as GDPR and detect unauthorized uses of personal data, we develop a new \emph{model auditing} technique that helps users check if their data was used to train a machine learning model. We focus on auditing deep-learning models that generate natural-language text, including word prediction and dialog generation. These models are at the core of popular online services and are often trained on personal data such as users' messages, searches, chats, and comments. We design and evaluate a black-box auditing method that can detect, with very few queries to a model, if a particular user's texts were used to train it (among thousands of other users). We empirically show that our method can successfully audit well-generalized models that are not overfitted to the training data. We also analyze how text-generation models memorize word sequences and explain why this memorization makes them amenable to auditing.
References in corpus (10)
- Google's Neural Machine Translation System: Bridging the Gap between Human and Machine Translation
- Learning Differentially Private Recurrent Language Models
- The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networks
- Differentially Private Federated Learning: A Client Level Perspective
- Chameleons in imagined conversations: A new approach to understanding coordination of linguistic style in dialogs
- On the importance of single directions for generalization
- Personalizing Dialogue Agents: I have a dog, do you have pets too?
- Understanding Membership Inferences on Well-Generalized Learning Models
- Towards Demystifying Membership Inference Attacks
- Stealing Hyperparameters in Machine Learning