A Scaling Law for Synthetic-to-Real Transfer: How Much Is Your Pre-training Effective?
This work addresses the challenge of optimizing synthetic data usage in vision tasks to reduce human labeling costs, though it is incremental as it builds on existing transfer learning frameworks.
The study tackled the problem of predicting performance in synthetic-to-real transfer learning by deriving a scaling law that estimates performance based on pre-training data amount, enabling decisions on whether to increase data or adjust synthesis settings, and validated it empirically across various benchmarks and model sizes.
Synthetic-to-real transfer learning is a framework in which a synthetically generated dataset is used to pre-train a model to improve its performance on real vision tasks. The most significant advantage of using synthetic images is that the ground-truth labels are automatically available, enabling unlimited expansion of the data size without human cost. However, synthetic data may have a huge domain gap, in which case increasing the data size does not improve the performance. How can we know that? In this study, we derive a simple scaling law that predicts the performance from the amount of pre-training data. By estimating the parameters of the law, we can judge whether we should increase the data or change the setting of image synthesis. Further, we analyze the theory of transfer learning by considering learning dynamics and confirm that the derived generalization bound is consistent with our empirical findings. We empirically validated our scaling law on various experimental settings of benchmark tasks, model sizes, and complexities of synthetic images.