Summary

This repository is a curated, continuously updated list of academic resources addressing the cold-start recommendation (CSR) problem. It aims to provide a comprehensive literature base for researchers and practitioners, with a particular focus on emerging solutions involving Large Language Models (LLMs).

How It Works

The repository organizes research papers into a detailed taxonomy, categorizing approaches such as data incomplete learning, knowledge alignment, meta-learning, graph-based methods, domain information transfer, and LLM integration. Papers are sorted chronologically, with direct links to publications and, where available, associated code repositories, facilitating quick access to relevant methodologies and enabling reproducibility.

Quick Start & Requirements

This is a curated list of academic resources, not a software project. It requires no installation and serves solely as a reference guide to research papers.

Highlighted Details

• LLM Integration Focus: A prominent trend is the application of LLMs for CSR, explored as direct recommenders, for prompt-based tuning, and as knowledge enhancers for representation and relation augmentation.
• Diverse Methodologies: Covers a broad spectrum of techniques including meta-learning, graph neural networks, contrastive learning, and generative models to address data scarcity.
• Structured Taxonomy: Research is systematically categorized into themes like "Data Incomplete Learning," "Graph Relations," and "Domain Information Transfer," offering a structured overview of the CSR landscape.
• Practical Resources: Many entries link directly to papers and their corresponding code implementations, supporting reproducibility and practical adoption.

Maintenance & Community

The repository is actively maintained and continuously updated, welcoming community contributions via issues or pull requests to include missing resources or new papers, serving the broader CSR research and industrial community.

Licensing & Compatibility

As a collection of links to academic papers and code, the repository itself does not have a specific license. Users must refer to the individual licenses of linked publications and codebases.

Limitations & Caveats

This list's comprehensiveness relies on community input and maintainer scope, focusing primarily on academic research. It may not cover all industry-specific solutions. The inclusion of pre-print papers (e.g., Arxiv 2025) means some content may not have undergone full peer review.