Timezone: »

Regularizing Neural Networks with Meta-Learning Generative Models
Shin'ya Yamaguchi · Daiki Chijiwa · Sekitoshi Kanai · Atsutoshi Kumagai · Hisashi Kashima

Wed Dec 13 03:00 PM -- 05:00 PM (PST) @ Great Hall & Hall B1+B2 #609

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 can be caused by the synthetic samples not perfectly representing class categories in real data and uniform sampling not necessarily providing 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 for regularizing feature extractors instead of training classifiers. These synthetic samples are dynamically determined to minimize the validation losses through meta-learning. We observed that MGR can avoid the performance degradation of naive 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 by up to 7 percentage points on test accuracy.

Author Information

Shin'ya Yamaguchi (NTT)
Daiki Chijiwa (Nippon Telegraph and Telephone Corporation)
Sekitoshi Kanai (NTT)
Atsutoshi Kumagai (NTT)
Hisashi Kashima (Kyoto University/RIKEN Center for AIP)

More from the Same Authors