# What makes instance discrimination good for transfer learning?

Source: [What makes instance discrimination good for transfer learning?](<https://www.sota2.com/research/paper/68fa691c1e209131cea204ba>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-27T06:05:44.223Z

Authors: Nanxuan Zhao, Zhirong Wu, Rynson W.H. Lau, Stephen Lin

Published: 2020-06-11T16:55:07.000Z

## Abstract

Contrastive visual pretraining based on the instance discrimination pretext task has made significant progress. Notably, recent work on unsupervised pretraining has shown to surpass the supervised counterpart for finetuning downstream applications such as object detection and segmentation. It comes as a surprise that image annotations would be better left unused for transfer learning. In this work, we investigate the following problems: What makes instance discrimination pretraining good for transfer learning? What knowledge is actually learned and transferred from these models? From this understanding of instance discrimination, how can we better exploit human annotation labels for pretraining? Our findings are threefold. First, what truly matters for the transfer is low-level and mid-level representations, not high-level representations. Second, the intra-category invariance enforced by the traditional supervised model weakens transferability by increasing task misalignment. Finally, supervised pretraining can be strengthened by following an exemplar-based approach without explicit constraints among the instances within the same category.

## 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.

- [Semantic Segmentation on PASCAL VOC 2012 (val)](<https://www.sota2.com/research/sota/semantic-segmentation-on-pascal-voc-2012-val.md>): Semantic segmentation; PASCAL VOC 2012 (val); Mean IoU: 76.52.

- [Image Classification on ImageNet 1k (val)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-val.md>): Image Classification; ImageNet-1k (val); Top-1 Accuracy: 70.1.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/2006.06606>)

[PDF](<https://arxiv.org/pdf/2006.06606.pdf>)
