RIM: Reliable Influence-based Active Learning on Graphs.

This repository is the official implementation of RIM.

Requirements

To install requirements:

pip install -r requirements.txt

Training

To train the model(s) in the paper:

cd the “example” data

run the python file RIM.py

Results

1. Accuracy comparison:

2. GCN performance comparison:

3. LP performance comparison:

4. Budget performance comparison:

5. Efficiency comparison:

6. Interpretability: