add pseudo-code

Author: cptq

README.md

@@ -7,7 +7,7 @@ Codebase for neural networks (SignNet and BasisNet) and experiments in the paper
 
 <img src="large_thumbnail.png" width=50%>
 
-### Experiments
+## Experiments
 
 `Alchemy` contains the experiments for graph-level regression on Alchemy.
 
@@ -17,14 +17,17 @@ Codebase for neural networks (SignNet and BasisNet) and experiments in the paper
 
 The intrinsic neural fields experiments use private code from the authors of the [original paper](https://arxiv.org/abs/2203.07967), so we do not yet publically release the SignNet codes for these.
 
-### Implementations
+## Implementations
 
 PyTorch Geometric SignNet for graph prediction: in `Alchemy`.
 
 DGL SignNet for graph prediction: in `GraphPrediction`.
 
 BasisNet for single graphs: in `LearningFilters`.
 
+The SignNet architecture is rather simple. Here is an example of pseudo-code for SignNet, as used for graph prediction tasks with a GNN base model:  
+<img src="pseudo-code.png" width=50%>
+
 <img src="LPE_symmetries.png" width=50%>
 
 **Coming Soon:** More experiments and implementations of our models! This repo and our paper are still a work in progress.