COUPA: An Industrial Recommender System for Online to Offline Service Platforms

Sicong Xie,Binbin Hu,Fengze Li,Ziqi Liu,Zhiqiang Zhang,Wenliang Zhong,Jun Zhou

DOI: https://doi.org/10.1145/3539618.3591828

2023-04-25

Abstract:Aiming at helping users locally discovery retail services (e.g., entertainment and dinning), Online to Offline (O2O) service platforms have become popular in recent years, which greatly challenge current recommender systems. With the real data in Alipay, a feeds-like scenario for O2O services, we find that recurrence based temporal patterns and position biases commonly exist in our scenarios, which seriously threaten the recommendation effectiveness. To this end, we propose COUPA, an industrial system targeting for characterizing user preference with following two considerations: (1) Time aware preference: we employ the continuous time aware point process equipped with an attention mechanism to fully capture temporal patterns for recommendation. (2) Position aware preference: a position selector component equipped with a position personalization module is elaborately designed to mitigate position bias in a personalized manner. Finally, we carefully implement and deploy COUPA on Alipay with a cooperation of edge, streaming and batch computing, as well as a two-stage online serving mode, to support several popular recommendation scenarios. We conduct extensive experiments to demonstrate that COUPA consistently achieves superior performance and has potential to provide intuitive evidences for recommendation

Information Retrieval,Machine Learning

What problem does this paper attempt to address?

### Problems the Paper Aims to Solve

This paper aims to address the recommendation system issues in online-to-offline (O2O) service platforms. Specifically, the paper focuses on the following two core problems:

1. **Time-Aware Preference Modeling**: How to effectively capture users' periodic preferences over continuous time? For example, users might visit specific malls or enjoy specific cuisines (like hotpot) weekly or daily, thus requiring a method to capture these time patterns.
2. **Location Preference Modeling**: How to handle users' click location preferences and establish a personalized debiasing model? Since the recommendation interface is usually displayed in the form of an information stream, users' attention distribution varies with different screen positions, bringing about personalized location bias.

To address these issues, the authors propose a new system called COUPA (COntinUous time and Position Aware recommender system). The system achieves its goals through the following methods:

- **Time-Aware Point Process Module**: Utilizes time function encoding to map timestamps into high-dimensional space and introduces a time-aware attention mechanism to capture the temporal relationships of important items.
- **Position Selector Component and Position Personalization Module**: Designs a position selector to mitigate position bias and models personalized location preferences through a Multi-gate Mixture-of-Experts (MMoE) network.
- **System Deployment and Implementation Details**: To support real-time user behavior capture and low-latency service, the system design includes technologies such as edge computing, stream computing, and batch processing.

In summary, by proposing the COUPA system, this paper aims to improve the performance of recommendation systems in O2O service platforms (such as Alipay), particularly under the dual influence of time and location preferences, to provide more accurate recommendation results.