GORAG: Graph-based Online Retrieval Augmented Generation for Dynamic Few-shot Social Media Text Classification

TL;DR

GORAG is a novel graph-based framework that enhances dynamic few-shot social media text classification by effectively retrieving and utilizing relevant contextual information, outperforming existing methods.

Contribution

It introduces a graph-based retrieval mechanism with edge weighting and dynamic context retrieval for improved few-shot text classification.

Findings

Outperforms existing approaches in dynamic few-shot settings.

Provides more comprehensive and precise contextual information.

Demonstrates effectiveness on social media text classification tasks.

Abstract

Text classification is vital for Web for Good applications like hate speech and misinformation detection. However, traditional models (e.g., BERT) often fail in dynamic few-shot settings where labeled data are scarce, and target labels frequently evolve. While Large Language Models (LLMs) show promise in few-shot settings, their performance is often hindered by increased input size in dynamic evolving scenarios. To address these issues, we propose GORAG, a Graph-based Online Retrieval-Augmented Generation framework for dynamic few-shot text classification. GORAG constructs and maintains a weighted graph of keywords and text labels, representing their correlations as edges. To model these correlations, GORAG employs an edge weighting mechanism to prioritize the importance and reliability of extracted information and dynamically retrieves relevant context using a tailored minimum-cost spanning tree for each input. Empirical evaluations show GORAG outperforms existing approaches by providing more comprehensive and precise contextual information. Our code is released at: https://github.com/Wyb0627/GORAG.