Field-Aware RankMixer with Dual-Stream Bilinear Fusion for the Tencent UNI-REC Challenge
Machine LearningArtificial Intelligence
Summary
The authors participated in a contest where the goal was to predict how likely users are to click and convert on ads by looking at their past actions across different types of activities and other related data. They built a new model called Field-Aware RankMixer that combines information from users' recent and older behaviors using a special technique to understand what interests them. Their model mixes different features using a method that looks at interactions between them and uses two parallel processes to improve prediction. This approach helped them place ninth in the competition. Their code is publicly available for others to explore.
Authors
Yufeng Zhang, Zhengqi Xu, Jiajun Cui
Abstract
This paper presents our solution to the KDD Cup 2026 Tencent UNIREC Challenge. The task requires joint modeling of multi-domain user behavior sequences and non-sequential multi-field features for target-ad pCVR prediction. We develop a Field-Aware RankMixer (FA-RankMixer) with dual-stream bilinear fusion. The model first applies target-aware DIN modules to extract user interests from multiple behavior domains. It also models recent and earlier interests separately for the longest behavior sequence. The model then forms semantic tokens based on feature fields and behavior domains and uses RankMixer blocks for cross-token interaction. A shallow MLP stream complements the deep RankMixer stream, and a group-wise bilinear module fuses their representations. Our final solution ranks ninth on the official leaderboard. Our code is available at https://github.com/PixelCookie-zyf/TAAC-2026-SeRankMixer.