Studies

Each study is a self-contained Python script located in the studies/ folder. Running a script generates numerical results (saved as .npy files) that are then consumed by the corresponding notebook for visualisation.

• Run All Studies
• Study A — Impact of \(p_e\) on convergence
• Study B — Optimal \(p_e\) for different \(N\)
• Study C — Impact of \(\alpha\) (learning rate)
• Study D — Optimal \(\alpha\) for different \(N\)
• Study E — Impact of \(p_e\) under larger batch size
• Study F — Distribution of weights during training
• Study G — Joint impact of \(\alpha\), \(p_e\), and batch size for different \(N\)
• Study H — Product Node vs MLP: Sparse vs Full-Table Training