MidRank: Learning to rank based on subsequences

TL;DR

MidRank is a supervised learning to rank algorithm that leverages the structure of image subsequences to improve ranking accuracy and generalization across various image ranking tasks.

Contribution

It introduces a novel approach that learns from moderately sized subsequences, enhancing learnability and generalization compared to pairwise or list-wise methods.

Findings

Significantly improves ranking accuracy on multiple datasets

Effectively exploits structural information in image sequences

Outperforms existing ranking methods in experiments

Abstract

We present a supervised learning to rank algorithm that effectively orders images by exploiting the structure in image sequences. Most often in the supervised learning to rank literature, ranking is approached either by analyzing pairs of images or by optimizing a list-wise surrogate loss function on full sequences. In this work we propose MidRank, which learns from moderately sized sub-sequences instead. These sub-sequences contain useful structural ranking information that leads to better learnability during training and better generalization during testing. By exploiting sub-sequences, the proposed MidRank improves ranking accuracy considerably on an extensive array of image ranking applications and datasets.