RE-SORT: Removing Spurious Correlation in Multilevel Interaction for CTR Prediction

TL;DR

RE-SORT is a novel CTR prediction framework that effectively removes spurious correlations in multilevel feature interactions, improving model accuracy and generalization by focusing on true causal features.

Contribution

The paper introduces RE-SORT, a framework combining multilevel stacked recurrent structures with a spurious correlation elimination module to enhance CTR prediction.

Findings

Achieves state-of-the-art accuracy on four CTR datasets.

Demonstrates improved generalization by removing confounding spurious correlations.

Operates efficiently with high speed and scalability.

Abstract

Click-through rate (CTR) prediction is a critical task in recommendation systems, serving as the ultimate filtering step to sort items for a user. Most recent cutting-edge methods primarily focus on investigating complex implicit and explicit feature interactions; however, these methods neglect the spurious correlation issue caused by confounding factors, thereby diminishing the model's generalization ability. We propose a CTR prediction framework that REmoves Spurious cORrelations in mulTilevel feature interactions, termed RE-SORT, which has two key components. I. A multilevel stacked recurrent (MSR) structure enables the model to efficiently capture diverse nonlinear interactions from feature spaces at different levels. II. A spurious correlation elimination (SCE) module further leverages Laplacian kernel mapping and sample reweighting methods to eliminate the spurious correlations concealed within the multilevel features, allowing the model to focus on the true causal features. Extensive experiments conducted on four challenging CTR datasets and our production dataset demonstrate that the proposed method achieves state-of-the-art performance in both accuracy and speed. The utilized codes, models and dataset will be released at https://github.com/RE-SORT.