ExplainRec: Towards Explainable Multi-Modal Zero-Shot Recommendation with Preference Attribution and Large Language Models

2025/10/03 by Ma, Bo, Liu, LuYao, Hu, ZeHua +1
Computer Science · #Explainable Artificial Intelligence (XAI) #Multimodal Machine Learning Applications #Recommender Systems and Techniques

paper · doi:10.48550/arxiv.2511.14770

Abstract

Recent advances in Large Language Models (LLMs) have opened new possibilities for recommendation systems, though current approaches such as TALLRec face challenges in explainability and cold-start scenarios. We present ExplainRec, a framework that extends LLM-based recommendation capabilities through preference attribution, multi-modal fusion, and zero-shot transfer learning. The framework incorporates four technical contributions: preference attribution tuning for explainable recommendations, zero-shot preference transfer for cold-start users and items, multi-modal enhancement leveraging visual and textual content, and multi-task collaborative optimization. Experimental evaluation on MovieLens-25M and Amazon datasets shows that ExplainRec outperforms existing methods, achieving AUC improvements of 0.7% on movie recommendation and 0.9% on cross-domain tasks, while generating interpretable explanations and handling cold-start scenarios effectively.

Citations

• Exploring the Upper Limits of Text-Based Collaborative Filtering Using Large Language Models: Discoveries and Insights
• Zero-Shot Next-Item Recommendation using Large Pretrained Language Models
• Chat-REC: Towards Interactive and Explainable LLMs-Augmented Recommender System
• LLaMA: Open and Efficient Foundation Language Models
• Recommendation as Language Processing (RLP): A Unified Pretrain, Personalized Prompt & Predict Paradigm (P5)
• Personalized Prompt Learning for Explainable Recommendation
• Personalized Prompt Learning for Explainable Recommendation
• LoRA: Low-Rank Adaptation of Large Language Models
• Cross-Domain Recommendation: Challenges, Progress, and Prospects
• Learning Transferable Visual Models From Natural Language Supervision
• LightGCN: Simplifying and Powering Graph Convolution Network for Recommendation
• LightGCN
• Sequential Scenario-Specific Meta Learner for Online Recommendation
• BERT4Rec: Sequential Recommendation with Bidirectional Encoder Representations from Transformer
• Self-Attentive Sequential Recommendation
• Attention Is All You Need
• Session-based Recommendations with Recurrent Neural Networks
• VBPR: Visual Bayesian Personalized Ranking from Implicit Feedback
• Image-based Recommendations on Styles and Substitutes
• Adam: A Method for Stochastic Optimization
• A Survey on Transfer Learning
• BPR: Bayesian Personalized Ranking from Implicit Feedback
• BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding

Related

• LLM4Rec: Large Language Models for Multimodal Generative Recommendation with Causal Debiasing
• Unlocking the Potential of Large Language Models for Explainable Recommendations
• LANE: Logic Alignment of Non-tuning Large Language Models and Online Recommendation Systems for Explainable Reason Generation
• Large Language Models as Narrative-Driven Recommenders
• User Preferences for Large Language Model versus Template-Based Explanations of Movie Recommendations: A Pilot Study