Reproduce “Toward User Preference Alignment in LLM Recommendation via Explicit Context Feedback” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Weizhi Zhang, Wooseong Yang, Yuxin Cui, Zhaohui Guo, Hins Hu, Liangwei Yang, Henry Peng Zou, Qifei Wang and 4 more
This paper proposes integrating explicit user feedback, such as reviews and comments, into Large Language Model (LLM)-based recommender systems. It outlines a multi-stage framework that leverages LLMs' natural language understanding to build detailed user profiles and enable context-aware retrieval and reranking, thereby improving preference alignment and transparency.
Key insights
- Explicit user feedback is a critical, underutilized resource containing nuanced rationales and context that can significantly enhance user preference alignment and mitigate filter bubbles.
- Large Language Models are uniquely positioned to process and interpret natural language explicit feedback, transforming it into actionable insights for personalized recommendations.
- New evaluation benchmarks and metrics, such as 'dislike avoidance rate' and 'preference fulfillment score,' are essential for adequately assessing context-aware recommender systems.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.