Representation Learning by Ranking across multiple tasks

TL;DR

This paper proposes a unified ranking-based framework for representation learning across diverse tasks, demonstrating its effectiveness in supervised and self-supervised settings with various data augmentation techniques.

Contribution

It introduces a novel perspective that formulates representation learning as a ranking problem, enabling a unified approach for multiple learning tasks.

Findings

Outperforms traditional methods in classification, retrieval, and multi-label tasks.

Shows significant advantages in self-supervised learning with data augmentation.

Validates the effectiveness of the ranking framework across various experimental setups.

Abstract

In recent years, representation learning has become the research focus of the machine learning community. Large-scale neural networks are a crucial step toward achieving general intelligence, with their success largely attributed to their ability to learn abstract representations of data. Several learning fields are actively discussing how to learn representations, yet there is a lack of a unified perspective. We convert the representation learning problem under different tasks into a ranking problem. By adopting the ranking problem as a unified perspective, representation learning tasks can be solved in a unified manner by optimizing the ranking loss. Experiments under various learning tasks, such as classification, retrieval, multi-label learning, and regression, prove the superiority of the representation learning by ranking framework. Furthermore, experiments under self-supervised learning tasks demonstrate the significant advantage of the ranking framework in processing unsupervised training data, with data augmentation techniques further enhancing its performance.