Timezone: »

 
The Trade-off between Label Efficiency and Universality of Representations from Contrastive Learning
Zhenmei Shi · Zhenmei Shi · Jiefeng Chen · Jiefeng Chen · Kunyang Li · Kunyang Li · Jayaram Raghuram · Jayaram Raghuram · Xi Wu · Xi Wu · Yingyiu Liang · Yingyiu Liang · Somesh Jha · Somesh Jha
Event URL: https://openreview.net/forum?id=oHoAwwQuVO_ »

The pre-train representation learning paradigm is a recent popular approach to address distribution shift and dataset limitation. This approach first pre-trains a representation function using large unlabeled datasets by self-supervised learning (e.g., contrastive learning), and then learns a classifier on the representation using small labeled datasets for downstream target tasks. The representation should have two key properties: label efficiency (i.e., learning an accurate classifier with a small amount of labeled data) and universality (i.e., useful for a wide range of downstream tasks). In this paper, we focus on contrastive learning and systematically study a trade-off between label efficiency and universality both empirically and theoretically. We empirically show that the trade-off exists in different models and datasets. Theoretically, we propose a data model with hidden representation and provide analysis in a simplified setting with linear models. The analysis shows that compared with pre-training on the target data directly, pre-training on diverse tasks can lead to a larger sample complexity for learning the classifier and thus worse prediction performance.

Author Information

Zhenmei Shi (UW-Madison)
Zhenmei Shi (UW-Madison)
Jiefeng Chen (University of Wisconsin-Madison)
Jiefeng Chen (University of Wisconsin-Madison)
Kunyang Li (University of Wisconsin - Madison)
Kunyang Li (University of Wisconsin - Madison)
Jayaram Raghuram (University of Wisconsin, Madison)
Jayaram Raghuram (University of Wisconsin, Madison)
Xi Wu (Google)

Completed my PhD in Computer Science from UW-Madison, advised by Jeffrey F. Naughton and Somesh Jha. Now a software engineer at Google. [Google PhD Fellow 2016 in privacy and security](https://ai.googleblog.com/2016/03/announcing-2016-google-phd-fellows-for.html).

Xi Wu (Google)

Completed my PhD in Computer Science from UW-Madison, advised by Jeffrey F. Naughton and Somesh Jha. Now a software engineer at Google. [Google PhD Fellow 2016 in privacy and security](https://ai.googleblog.com/2016/03/announcing-2016-google-phd-fellows-for.html).

Yingyiu Liang (University of Wisconsin-Madison)
Yingyiu Liang (University of Wisconsin-Madison)
Somesh Jha (University of Wisconsin, Madison)
Somesh Jha (University of Wisconsin, Madison)

More from the Same Authors