Abstract:Multi-domain recommendation (MDR) aims to enhance recommendation performance across various domains. However, real-world recommender systems in online platforms often need to handle dozens or even hundreds of domains, far exceeding the capabilities of traditional MDR algorithms, which typically focus on fewer than five domains. Key challenges include a substantial increase in parameter count, high maintenance costs, and intricate knowledge transfer patterns across domains. Furthermore, minor domains often suffer from data sparsity, leading to inadequate training in classical methods. To address these issues, we propose Adaptive REcommendation for All Domains with counterfactual augmentation (AREAD). AREAD employs a hierarchical structure with a limited number of expert networks at several layers, to effectively capture domain knowledge at different granularities. To adaptively capture the knowledge transfer pattern across domains, we generate and iteratively prune a hierarchical expert network selection mask for each domain during training. Additionally, counterfactual assumptions are used to augment data in minor domains, supporting their iterative mask pruning. Our experiments on two public datasets, each encompassing over twenty domains, demonstrate AREAD's effectiveness, especially in data-sparse domains. Source code is available at <a class="link-external link-https" href="https://github.com/Chrissie-Law/AREAD-Multi-Domain-Recommendation" rel="external noopener nofollow">this https URL</a>.

What problem does this paper attempt to address?

This paper attempts to solve several key problems faced by multi - domain recommendation systems (Multi - Domain Recommendation, MDR) when dealing with a large number of domains. Specifically, these problems include: 1. **System scalability issues**: Traditional MDR algorithms can usually handle only a few domains (usually less than 5), but in the real world, recommendation systems need to handle dozens or even hundreds of domains. This leads to a significant increase in the number of model parameters, high maintenance costs, and complex cross - domain knowledge transfer patterns. In addition, the data sparsity problem in small domains also results in insufficient training. 2. **Complex cross - domain knowledge transfer patterns**: As the number of domains increases, the traditional binary classification framework that divides knowledge into domain - shared or domain - specific becomes no longer applicable. Knowledge transfer between different domains is directional, and the knowledge of some domains may not be applicable to other domains, so determining which domains should be learned together becomes a challenge. 3. **Large differences in sample sizes across domains**: As shown in Figure 1, nearly half of the domain sample sizes are less than 2% of the total sample size. This huge difference in sample sizes causes existing MDR models to be dominated by data - rich domains during the optimization process, thus affecting the optimization effect on data - sparse domains (especially small domains). To solve the above problems, the authors propose a new framework named Adaptive REcommendation for All Domains with counterfactual augmentation (AREAD). AREAD addresses these challenges in the following ways: - **Hierarchical Expert Integration (HEI)**: Use a small number of expert networks in the hierarchical structure to capture domain knowledge at different granularities, reduce the number of parameters, and improve the scalability of the model. - **Hierarchical Expert Mask Pruning (HEMP)**: Select specific experts for each domain by iteratively generating and pruning masks to adapt to complex cross - domain knowledge transfer patterns. - **Popularity - based Counterfactual Augmentation**: Perform counterfactual augmentation on the data of small domains to compensate for the data sparsity problem. The experimental results show that AREAD performs excellently on two public datasets (each dataset contains more than 20 domains), especially with a significant improvement in data - sparse domains.

One for Dozens: Adaptive REcommendation for All Domains with Counterfactual Augmentation

Adaptive Domain Interest Network for Multi-domain Recommendation

MAMDR: A Model Agnostic Learning Method for Multi-Domain Recommendation

Multi-domain Recommendation with Embedding Disentangling and Domain Alignment

Cross-domain recommendation with Multi-Auxiliary Domains via Consistent and Selective Cluster-Level Knowledge Transfer

Cross-domain Recommendation Via Dual Adversarial Adaptation

Cross-domain collaborative recommendation without overlapping entities based on domain adaptation

MultiLoRA: Multi-Directional Low Rank Adaptation for Multi-Domain Recommendation

Multi-Domain Recommendation to Attract Users via Domain Preference Modeling

Cross-domain Recommendation Via Adversarial Adaptation.

HAMUR: Hyper Adapter for Multi-Domain Recommendation

D3: A Methodological Exploration of Domain Division, Modeling, and Balance in Multi-Domain Recommendations

One for All, All for One: Learning and Transferring User Embeddings for Cross-Domain Recommendation

Multi-Sparse-Domain Collaborative Recommendation via Enhanced Comprehensive Aspect Preference Learning

Diverse Preference Augmentation with Multiple Domains for Cold-start Recommendations

Beyond the Overlapping Users: Cross-Domain Recommendation via Adaptive Anchor Link Learning

DADIN: Domain Adversarial Deep Interest Network for cross domain recommender systems

AMT-CDR: A Deep Adversarial Multi-channel Transfer Network for Cross-domain Recommendation

Adaptive Conditional Expert Selection Network for Multi-domain Recommendation

Large-Scale Multi-Domain Recommendation: an Automatic Domain Feature Extraction and Personalized Integration Framework

Transferring Causal Mechanism over Meta-representations for Target-unknown Cross-domain Recommendation