EXIT: An EXplicit Interest Transfer Framework for Cross-Domain Recommendation

TL;DR

EXIT tackles cross-domain recommendation by addressing negative transfer through an explicit, supervised transfer mechanism. It introduces three components: an Interest Prediction Network (IPN) to model target and source domain interests, an Interest Combination Label (ICL) to supervise transfer via an explicit transfer probability, and a Scene Selector Network (SSN) to modulate transfer intensity in fine-grained contexts, enabling the complete target-domain interest to be expressed as $P_{whole} = P_{target} + P_{source} \cdot P_{trans}$. A multi-task joint loss ties together target/sources predictions with the transfer supervision, and online truncation ensures bounded outputs. Offline and online experiments on Meituan data show EXIT surpasses single-domain and existing cross-domain baselines, with measurable boosts in CTCVR and GTV and reduced NFR, leading to successful production deployment on Meituan’s homepage. This explicit paradigm reduces negative transfer and provides practical, scalable deployment for real-world cross-domain recommendations.

Abstract

Cross-domain recommendation has attracted substantial interest in industrial apps such as Meituan, which serves multiple business domains via knowledge transfer and meets the diverse interests of users. However, existing methods typically follow an implicit modeling paradigm that blends the knowledge from both the source and target domains, and design intricate network structures to share learned embeddings or patterns between domains to improve recommendation accuracy. Since the transfer of interest signals is unsupervised, these implicit paradigms often struggle with the negative transfer resulting from differences in service functions and presentation forms across different domains. In this paper, we propose a simple and effective EXplicit Interest Transfer framework named EXIT to address the stated challenge. Specifically, we propose a novel label combination approach that enables the model to directly learn beneficial source domain interests through supervised learning, while excluding inappropriate interest signals. Moreover, we introduce a scene selector network to model the interest transfer intensity under fine-grained scenes. Offline experiments conducted on the industrial production dataset and online A/B tests validate the superiority and effectiveness of our proposed framework. Without complex network structures or training processes, EXIT can be easily deployed in the industrial recommendation system. EXIT has been successfully deployed in the online homepage recommendation system of Meituan App, serving the main traffic.

Figures (4)

• Figure 1: Three representative business domains on Meituan mobile app: Homepage Recommendation, Channel Section, and Search. Each domain offers a variety of businesses, including hotel booking, food delivery, group deals, etc.
• Figure 2: High-level overview of the EXIT. Unlike traditional CDR methods, EXIT explicitly transfers interests from the source domain that are beneficial to the target domain based on user's specific contexts, preventing negative transfer.
• Figure 3: The framework of the EXIT. EXIT models the user's target domain interest $P_{target}$, the source domain interest $P_{source}$, and the cross-domain interest transfer probability $P_{trans}$ , ultimately obtaining the user’s complete interest suitable for the target domain as $P_{whole} = P_{target}+P_{source}*P_{trans}$.
• Figure 4: Parameter sensitivity of $\lambda_1,\lambda_2$ and $\lambda_3$.