On the Trade-off of Intra-/Inter-class Diversity for Supervised Pre-training

Jieyu Zhang · Bohan Wang · Zhengyu Hu · Pang Wei Koh · Alexander Ratner

Great Hall & Hall B1+B2 (level 1) #429
[ ]
Wed 13 Dec 8:45 a.m. PST — 10:45 a.m. PST


Pre-training datasets are critical for building state-of-the-art machine learning models, motivating rigorous study on their impact on downstream tasks. In this work, we study the impact of the trade-off between the intra-class diversity (the number of samples per class) and the inter-class diversity (the number of classes) of a supervised pre-training dataset. Empirically, we found that with the size of the pre-training dataset fixed, the best downstream performance comes with a balance on the intra-/inter-class diversity. To understand the underlying mechanism, we show theoretically that the downstream performance depends monotonically on both types of diversity. Notably, our theory reveals that the optimal class-to-sample ratio (#classes / #samples per class) is invariant to the size of the pre-training dataset, which motivates an application of predicting the optimal number of pre-training classes. We demonstrate the effectiveness of this application by an improvement of around 2 points on the downstream tasks when using ImageNet as the pre-training dataset.

Chat is not available.