AgentCF++: Memory-enhanced LLM-based Agents for Popularity-aware Cross-domain Recommendations
This work addresses cross-domain recommendation challenges for users influenced by popularity, though it is incremental in enhancing existing agent-based methods.
The paper tackles the problem of LLM-based user agents introducing irrelevant information and failing to capture popularity influences in cross-domain recommendations, proposing a dual-layer memory architecture and group-shared memory that achieve improvements, such as a 5.2% increase in recall@10 on a benchmark dataset.
LLM-based user agents, which simulate user interaction behavior, are emerging as a promising approach to enhancing recommender systems. In real-world scenarios, users' interactions often exhibit cross-domain characteristics and are influenced by others. However, the memory design in current methods causes user agents to introduce significant irrelevant information during decision-making in cross-domain scenarios and makes them unable to recognize the influence of other users' interactions, such as popularity factors. To tackle this issue, we propose a dual-layer memory architecture combined with a two-step fusion mechanism. This design avoids irrelevant information during decision-making while ensuring effective integration of cross-domain preferences. We also introduce the concepts of interest groups and group-shared memory to better capture the influence of popularity factors on users with similar interests. Comprehensive experiments validate the effectiveness of AgentCF++. Our code is available at https://github.com/jhliu0807/AgentCF-plus.