Cross-Domain Recommendation Meets Large Language Models
Ajay Krishna Vajjala, Dipak Meher, Ziwei Zhu, David S. Rosenblum
TL;DR
This paper investigates the use of large language models (LLMs) for cross-domain recommendation (CDR) to address cold-start issues in single-domain systems. It introduces two prompt designs—target-domain behavior injection and no target-domain behavior injection—to steer LLMs for rating and ranking in CDR and evaluates them across three domain pairs using multiple LLMs and baselines. Across intuitive domain pairs, LLMs, especially GPT-4 family and GPT-4o, tend to outperform state-of-the-art baselines on ranking and rating tasks, with high-context prompts generally providing the best performance; dissimilar domain pairs remain challenging. The work highlights the potential of LLMs as cross-domain recommenders, emphasizes the importance of prompt design and model scale, and points to future directions such as hybrid models and adaptive prompts for broader CDR applicability.
Abstract
Cross-domain recommendation (CDR) has emerged as a promising solution to the cold-start problem, faced by single-domain recommender systems. However, existing CDR models rely on complex neural architectures, large datasets, and significant computational resources, making them less effective in data-scarce scenarios or when simplicity is crucial. In this work, we leverage the reasoning capabilities of large language models (LLMs) and explore their performance in the CDR domain across multiple domain pairs. We introduce two novel prompt designs tailored for CDR and demonstrate that LLMs, when prompted effectively, outperform state-of-the-art CDR baselines across various metrics and domain combinations in the rating prediction and ranking tasks. This work bridges the gap between LLMs and recommendation systems, showcasing their potential as effective cross-domain recommenders.