# Why Fine-grained Labels in Pretraining Benefit Generalization?

arXiv: https://arxiv.org/abs/2410.23129 · HTML: https://mlsanity.com/paper/2410.23129/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 30 Oct 2024.

## Built on by (most cited first, 3 of 3)

- [Perch 2.0: The Bittern Lesson for Bioacoustics](https://mlsanity.com/paper/2508.04665/) (2025)
- [Perch 2.0 transfers 'whale' to underwater tasks](https://mlsanity.com/paper/2512.03219/) (2025)
- [The Advantage of Fine-Grained Training](https://mlsanity.com/paper/2509.05130/) (2025)

## Abstract (arXiv)

Recent studies show that pretraining a deep neural network with fine-grained labeled data, followed by fine-tuning on coarse-labeled data for downstream tasks, often yields better generalization than pretraining with coarse-labeled data. While there is ample empirical evidence supporting this, the theoretical justification remains an open problem. This paper addresses this gap by introducing a "hierarchical multi-view" structure to confine the input data distribution. Under this framework, we prove that: 1) coarse-grained pretraining only allows a neural network to learn the common features well, while 2) fine-grained pretraining helps the network learn the rare features in addition to the common ones, leading to improved accuracy on hard downstream test samples.