TranSUN: A Preemptive Paradigm to Eradicate Retransformation Bias Intrinsically from Regression Models in Recommender Systems

TL;DR

This paper introduces TranSUN, a preemptive regression model paradigm that intrinsically eliminates retransformation bias in recommender systems, outperforming existing post-hoc correction methods and demonstrating real-world industrial deployment.

Contribution

The paper proposes a novel bias-free regression framework, TranSUN, with a joint bias learning approach, and generalizes it into GTS, enabling flexible development of unbiased models in recommender systems.

Findings

TranSUN achieves superior unbiasedness and convergence in experiments.

The methods outperform existing bias correction techniques.

Successful deployment in Taobao's product and video recommendation systems.

Abstract

Regression models are crucial in recommender systems. However, retransformation bias problem has been conspicuously neglected within the community. While many works in other fields have devised effective bias correction methods, all of them are post-hoc cures externally to the model, facing practical challenges when applied to real-world recommender systems. Hence, we propose a preemptive paradigm to eradicate the bias intrinsically from the models via minor model refinement. Specifically, a novel TranSUN method is proposed with a joint bias learning manner to offer theoretically guaranteed unbiasedness under empirical superior convergence. It is further generalized into a novel generic regression model family, termed Generalized TranSUN (GTS), which not only offers more theoretical insights but also serves as a generic framework for flexibly developing various bias-free models. Comprehensive experimental results demonstrate the superiority of our methods across data from various domains, which have been successfully deployed in two real-world industrial recommendation scenarios, i.e. product and short video recommendation scenarios in Guess What You Like business domain in the homepage of Taobao App (a leading e-commerce platform with DAU > 300M), to serve the major online traffic.