English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Recommender systems are prone to be misled by biases in the data. Modelstrained with biased data fail to capture the real interests of users, thus itis critical to alleviate the impact of bias to achieve unbiased recommendation.In this work, we focus on an essential bias in micro-video recommendation,duration bias. Specifically, existing micro-video recommender systems usuallyconsider watch time as the most critical metric, which measures how long a userwatches a video. Since videos with longer duration tend to have longer watchtime, there exists a kind of duration bias, making longer videos tend to berecommended more against short videos. In this paper, we empirically show thatcommonly-used metrics are vulnerable to duration bias, making them NOT suitablefor evaluating micro-video recommendation. To address it, we further propose anunbiased evaluation metric, called WTG (short for Watch Time Gain). Empiricalresults reveal that WTG can alleviate duration bias and better measurerecommendation performance. Moreover, we design a simple yet effective modelnamed DVR (short for Debiased Video Recommendation) that can provide unbiasedrecommendation of micro-videos with varying duration, and learn unbiased userpreferences via adversarial learning. Extensive experiments based on tworeal-world datasets demonstrate that DVR successfully eliminates duration biasand significantly improves recommendation performance with over 30% relativeprogress. Codes and datasets are released athttps://github.com/tsinghua-fib-lab/WTG-DVR.

DVR: Micro-Video Recommendation Optimizing Watch-Time-Gain under Duration Bias