Data

Tools

Indexes

Applications

Experiments

Open Source Science

nn-simulation

https://github.com/thibaudjoel/nn-simulation

Science Score: 31.0%

This score indicates how likely this project is to be science-related based on various indicators:

  • ✓
    CITATION.cff file
    Found CITATION.cff file
  • ✓
    codemeta.json file
    Found codemeta.json file
  • ○
    .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (9.5%) to scientific vocabulary
Last synced: 10 months ago · JSON representation ·

Repository

Basic Info
  • Host: GitHub
  • Owner: thibaudjoel
  • Language: Jupyter Notebook
  • Default Branch: main
  • Size: 153 KB
Statistics
  • Stars: 0
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created over 1 year ago · Last pushed about 1 year ago
Metadata Files
Readme Citation

README.md

nn-simulation

This repository contains the code used for the numerical experiments in my Master thesis:

“Non-asymptotic guarantees for parameter inference in neural networks”

Repository Overview

Core Modules

  • utils.py

    • Helper functions to compute the activation function $\sigma$ and its derivatives.
    • Helper function to set up folder structure for experiment results.

  • Convergence.py

    • Simulates Multinomial data.
    • Tracks the convergence of the MLE procedure for the neural network model.
    • Saves results as .json.

  • MonteCarlo.py

    • Runs Monte Carlo simulations for MLE procedures with various parameters.
    • Stores outcomes in .json format.

  • DataClasses.py

    • Loads experiment data from JSON files.
    • Visualizes results, with options to compare parameters and export plots.

Notebooks

  • experiments.ipynb

    • Sets up folder structure.
    • Runs both convergence and Monte Carlo simulations.

  • plots.ipynb

    • Generates plots based on experimental results.

Folder Structure

Important:
Run create_folders() from utils.py before executing any simulations. Otherwise, the code will fail when trying to save output files. This will create the following directory structure in the current working directory:

```text data/ ├── conv/ # Convergence experiment results ├── mc/ # Monte Carlo simulation results

imgs/ ├── conv/ # Convergence plots ├── mc/ # Monte Carlo plots ```

Parallel Execution

The code is designed to run in parallel for faster computation.
It is highly recommended to use a machine with multiple CPU cores to speed up processing time.

Owner

  • Login: thibaudjoel
  • Kind: user

Citation (CITATION.cff) 

cff-version: 1.2.0
authors:
  - family-names: Hadamczik
    given-names: Thibaud Joel
title: "Neural Network Simulation"
url: "https://github.com/thibaudjoel/nn-simulation"

GitHub Events

Total
  • Push event: 42
  • Create event: 2
Last Year
  • Push event: 42
  • Create event: 2