# Regularizing Neural Networks with Meta-Learning Generative Models Source: [Regularizing Neural Networks with Meta-Learning Generative Models]() Collection index: [llms.txt]() Data updated: 2026-01-04T04:37:35.589Z Authors: Shin'ya Yamaguchi, Daiki Chijiwa, Sekitoshi Kanai, Atsutoshi Kumagai, Hisashi Kashima Published: 2023-07-26T01:47:49.000Z ## Abstract This paper investigates methods for improving generative data augmentation for deep learning. Generative data augmentation leverages the synthetic samples produced by generative models as an additional dataset for classification with small dataset settings. A key challenge of generative data augmentation is that the synthetic data contain uninformative samples that degrade accuracy. This is because the synthetic samples do not perfectly represent class categories in real data and uniform sampling does not necessarily provide useful samples for tasks. In this paper, we present a novel strategy for generative data augmentation called meta generative regularization (MGR). To avoid the degradation of generative data augmentation, MGR utilizes synthetic samples in the regularization term for feature extractors instead of in the loss function, e.g., cross-entropy. These synthetic samples are dynamically determined to minimize the validation losses through meta-learning. We observed that MGR can avoid the performance degradation of na\\"ive generative data augmentation and boost the baselines. Experiments on six datasets showed that MGR is effective particularly when datasets are smaller and stably outperforms baselines. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Image Classification on Stanford Cars 10% subset](): Image Classification; Stanford Cars 10% subset; Top-1 Acc: 23.49. - [Image Classification on Stanford Cars (25% subset)](): Image Classification; Stanford Cars 25% subset; Top-1 Accuracy: 53.16. - [Image Classification on Stanford Cars 50%](): Image Classification; Stanford Cars 50%; Top-1 Accuracy: 75.24. - [Image Classification on Stanford Cars 75%](): Image Classification; Stanford Cars 75%; Top-1 Accuracy: 83.13. - [Classification on CIFAR-10-C 1.0 (test)](): Classification; CIFAR-10-C 1.0 (test); Accuracy (Clean): 88.02. - [Image Classification on ImageNet (val) (ACC)](): Image Classification; ImageNet (val); Accuracy: not recorded. ## Original sources and implementation [Original paper]() [PDF]()