Data-Copying in Generative Models: A Formal Framework
This addresses the issue of data-copying in generative models for researchers and practitioners, but it is incremental as it builds on prior work.
The paper tackles the problem of detecting memorization in generative models by showing that an existing framework fails to detect certain types of blatant memorization, and it provides an alternative definition and method for detection with provable guarantees and sample requirement lower bounds.
There has been some recent interest in detecting and addressing memorization of training data by deep neural networks. A formal framework for memorization in generative models, called "data-copying," was proposed by Meehan et. al. (2020). We build upon their work to show that their framework may fail to detect certain kinds of blatant memorization. Motivated by this and the theory of non-parametric methods, we provide an alternative definition of data-copying that applies more locally. We provide a method to detect data-copying, and provably show that it works with high probability when enough data is available. We also provide lower bounds that characterize the sample requirement for reliable detection.