This research project explores how Deep Neural Networks(DNN) can learn composition of functions with bounded F-1 Norm, referenced in [How DNNs Break the Curse of Dimensionality: Compositionality and Symmetry Learning]
This page includes practical implementations in Python with associated experimental results. These notebooks contain the code, data, and visualizations used in our study, allowing for reproducibility and further exploration.
notebooks/: Directory containing Jupyter notebooks with code and experiments.data/: Directory with both syntheic & real-world datasets used in our experimentsREADME.md: This file, providing an overview and instructions.
To run the code in this project, you will need to have Python and Jupyter installed. You can set up a virtual environment and install the required packages using the following commands:
# Clone the repository
git clone https://github.com/shc443/CoveringNumber_GB/edit/main/README.md
cd your-repo-name
# Install required packages
pip install -r requirements.txt