Aqueous Solubility Prediction

Overview

Applied graph neural network (SolTranNet) to predict aqueous solubility classification in a Kaggle competition. This project demonstrates the application of GNNs to molecular property prediction.

Technical Details

• Model: SolTranNet (Graph Neural Network)
• Framework: PyTorch
• Task: Binary classification of molecular solubility
• Competition: Kaggle

Approach

Leveraged graph neural networks to model molecular structures as graphs, where atoms are nodes and bonds are edges. The GNN learns to predict solubility based on molecular topology and atomic features.