Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Survey on Large Language Models for Personalized and Explainable Recommendations

Junyi Chen

TL;DR

This survey analyzes how Large Language Models can enhance personalization and explainability in Recommender Systems, with a focus on prompt-based and parameter-efficient strategies. It covers transformer-based NLG architectures, pre-training vs. prompting, and PEFT methods, as well as PEG-specific techniques like PEPLER and rationale/in-text learning, alongside evaluation metrics for quality and diversity. The paper also discusses PEG-related challenges such as cold-start and biases, highlighting approaches like PromptRec and bias-mitigating prompts to address them. Together, these insights offer a roadmap for building scalable, transparent, and fair LLM-enabled RS that can generate natural language explanations aligned with user needs.

Abstract

In recent years, Recommender Systems(RS) have witnessed a transformative shift with the advent of Large Language Models(LLMs) in the field of Natural Language Processing(NLP). These models such as OpenAI's GPT-3.5/4, Llama from Meta, have demonstrated unprecedented capabilities in understanding and generating human-like text. This has led to a paradigm shift in the realm of personalized and explainable recommendations, as LLMs offer a versatile toolset for processing vast amounts of textual data to enhance user experiences. To provide a comprehensive understanding of the existing LLM-based recommendation systems, this survey aims to analyze how RS can benefit from LLM-based methodologies. Furthermore, we describe major challenges in Personalized Explanation Generating(PEG) tasks, which are cold-start problems, unfairness and bias problems in RS.

A Survey on Large Language Models for Personalized and Explainable Recommendations

TL;DR

This survey analyzes how Large Language Models can enhance personalization and explainability in Recommender Systems, with a focus on prompt-based and parameter-efficient strategies. It covers transformer-based NLG architectures, pre-training vs. prompting, and PEFT methods, as well as PEG-specific techniques like PEPLER and rationale/in-text learning, alongside evaluation metrics for quality and diversity. The paper also discusses PEG-related challenges such as cold-start and biases, highlighting approaches like PromptRec and bias-mitigating prompts to address them. Together, these insights offer a roadmap for building scalable, transparent, and fair LLM-enabled RS that can generate natural language explanations aligned with user needs.

Abstract

In recent years, Recommender Systems(RS) have witnessed a transformative shift with the advent of Large Language Models(LLMs) in the field of Natural Language Processing(NLP). These models such as OpenAI's GPT-3.5/4, Llama from Meta, have demonstrated unprecedented capabilities in understanding and generating human-like text. This has led to a paradigm shift in the realm of personalized and explainable recommendations, as LLMs offer a versatile toolset for processing vast amounts of textual data to enhance user experiences. To provide a comprehensive understanding of the existing LLM-based recommendation systems, this survey aims to analyze how RS can benefit from LLM-based methodologies. Furthermore, we describe major challenges in Personalized Explanation Generating(PEG) tasks, which are cold-start problems, unfairness and bias problems in RS.
Paper Structure (13 sections, 9 equations, 6 figures)

This paper contains 13 sections, 9 equations, 6 figures.

Table of Contents

  1. Introduction
  2. Natural Language Generation
  3. Methodologies
  4. Transformer Architecture
  5. Training strategies
  6. Explanation Generating for Recommendation
  7. PEPLER
  8. Rationale Generation and In-text Learning
  9. Evaluation Metrics
  10. Challenges
  11. Cold-Start Problem
  12. Bias Problem
  13. Conclusion

Figures (6)

  • Figure 1: Architecture of Transformer
  • Figure 2: A generic architecture of language modeling paradigm for recommendation purpose
  • Figure 3: Four different training manners for LLM-based recommendations
  • Figure 4: the PEPLER-D method that utilizes item features as a discrete prompt
  • Figure 5: the PEPLER method that treats user and item IDs as continuous prompts
  • ...and 1 more figures