\textsc{Text2Graph}: Combining Lightweight LLMs and GNNs for Efficient Text Classification in Label-Scarce Scenarios
By: João Lucas Luz Lima Sarcinelli, Ricardo Marcondes Marcacini
Potential Business Impact:
Makes AI smarter using less energy.
Large Language Models (LLMs) have become effective zero-shot classifiers, but their high computational requirements and environmental costs limit their practicality for large-scale annotation in high-performance computing (HPC) environments. To support more sustainable workflows, we present \textsc{Text2Graph}, an open-source Python package that provides a modular implementation of existing text-to-graph classification approaches. The framework enables users to combine LLM-based partial annotation with Graph Neural Network (GNN) label propagation in a flexible manner, making it straightforward to swap components such as feature extractors, edge construction methods, and sampling strategies. We benchmark \textsc{Text2Graph} on a zero-shot setting using five datasets spanning topic classification and sentiment analysis tasks, comparing multiple variants against other zero-shot approaches for text classification. In addition to reporting performance, we provide detailed estimates of energy consumption and carbon emissions, showing that graph-based propagation achieves competitive results at a fraction of the energy and environmental cost.
Similar Papers
Court of LLMs: Evidence-Augmented Generation via Multi-LLM Collaboration for Text-Attributed Graph Anomaly Detection
Machine Learning (CS)
Finds weird patterns in text and connections.
Leveraging Language Semantics for Collaborative Filtering with TextGCN and TextGCN-MLP: Zero-Shot vs In-Domain Performance
Information Retrieval
Helps computers suggest movies you'll like.
Large Language Models Meet Text-Attributed Graphs: A Survey of Integration Frameworks and Applications
Computation and Language
Helps computers understand and reason better.