Alleviating Video-Length Effect for Micro-video Recommendation

TL;DR

This paper addresses the bias introduced by video length in micro-video recommendation systems, proposing a debiasing method that improves user engagement and content matching.

Contribution

The paper introduces VLDRec, a novel approach combining data labeling, sample generation, and multi-task learning to mitigate video length bias in recommendations.

Findings

VLDRec improves view time by 1.81% and 11.32% on two datasets.

It effectively reduces length bias in user preference modeling.

Enhances content relevance in recommendations.

Abstract

Micro-videos platforms such as TikTok are extremely popular nowadays. One important feature is that users no longer select interested videos from a set, instead they either watch the recommended video or skip to the next one. As a result, the time length of users' watching behavior becomes the most important signal for identifying preferences. However, our empirical data analysis has shown a video-length effect that long videos are easier to receive a higher value of average view time, thus adopting such view-time labels for measuring user preferences can easily induce a biased model that favors the longer videos. In this paper, we propose a Video Length Debiasing Recommendation (VLDRec) method to alleviate such an effect for micro-video recommendation. VLDRec designs the data labeling approach and the sample generation module that better capture user preferences in a view-time oriented manner. It further leverages the multi-task learning technique to jointly optimize the above samples with original biased ones. Extensive experiments show that VLDRec can improve the users' view time by 1.81% and 11.32% on two real-world datasets, given a recommendation list of a fixed overall video length, compared with the best baseline method. Moreover, VLDRec is also more effective in matching users' interests in terms of the video content.