Data

Tools

Indexes

Applications

Experiments

Open Source Science

inn_ijoc

code for the Integer Neural Networks on IJOC

https://github.com/ambrogiomb/inn_ijoc

Science Score: 44.0%

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
    Found .zenodo.json file
  • ○
    DOI references
  • ○
    Academic publication links
  • ○
    Academic email domains
  • ○
    Institutional organization owner
  • ○
    JOSS paper metadata
  • ○
    Scientific vocabulary similarity
    Low similarity (5.4%) to scientific vocabulary
Last synced: 6 months ago · JSON representation ·

Repository

code for the Integer Neural Networks on IJOC

Basic Info
  • Host: GitHub
  • Owner: AmbrogioMB
  • License: mit
  • Language: Python
  • Default Branch: main
  • Size: 43.9 KB
Statistics
  • Stars: 0
  • Watchers: 1
  • Forks: 0
  • Open Issues: 0
  • Releases: 0
Created over 2 years ago · Last pushed over 1 year ago
Metadata Files
Readme License Citation

README.md

INN_IJOC

Code for the article Multi-Objective Linear Ensembles for Robust and Sparse Training of Few-Bit Neural Networks by A. M. Bernardelli, S. Gualandi, H. C. Lau, S. Milanesi, N. Yorke-Smith.

The repository contains one python file only, that can be used to reproduce the tests regarding the MNIST dataset.

The PARAMETERS can be modified accordingly to the experiments.

For every value in the list different_p, the file produces 3nm + 1 csv files where n is the number of instances, namely instances, and m is the number of different numbers of training images, namely the length of the list training_images Fixed the training images i and the instance j, the file produces - labels_i_j.csv for every image used in the tests, how many networks labelled a certain image with a certain label - test_inn_i_j.csv for every network of the ensemble, some infos about gap, time, etc., are collected - test_inn_i_j_weights.csv for every network of the ensemble, the weights distribution is collected.

The last file, test_inn.csv, contains results on accuracy and label statuses.

Owner

  • Name: Ambrogio Maria Bernardelli
  • Login: AmbrogioMB
  • Kind: user

PhD student in Computational Mathematics at University of Pavia

Citation (CITATION.cff) 

cff-version: 1.2.0
message: "If you use this code, please cite it as below."
authors:
- family-names: "Bernardelli"
  given-names: "Ambrogio Maria"
  orcid: "https://orcid.org/0000-0002-2328-7062"
- family-names: "Milanesi"
  given-names: "Simone"
  orcid: "https://orcid.org/0000-0002-6314-1965"
- family-names: "Gualandi"
  given-names: "Stefano"
  orcid: "https://orcid.org/0000-0002-2111-3528"
- family-names: "Lau"
  given-names: "Hoong Chuin"
  orcid: "https://orcid.org/0000-0002-5326-411X"
- family-names: "Yorke-Smith"
  given-names: "Neil"
  orcid: "https://orcid.org/0000-0002-1814-3515"
title: "MILP Training of Few-Bit Neural Networks"
date-released: 2023-06-30
version: 0.1.0
url: "https://github.com/AmbrogioMB/INN_IJOC"
preferred-citation:
type: article
authors:
- family-names: "Bernardelli"
  given-names: "Ambrogio Maria"
  orcid: "https://orcid.org/0000-0002-2328-7062"
- family-names: "Milanesi"
  given-names: "Simone"
  orcid: "https://orcid.org/0000-0002-6314-1965"
- family-names: "Gualandi"
  given-names: "Stefano"
  orcid: "https://orcid.org/0000-0002-2111-3528"
- family-names: "Lau"
  given-names: "Hoong Chuin"
  orcid: "https://orcid.org/0000-0002-5326-411X"
- family-names: "Yorke-Smith"
  given-names: "Neil"
  orcid: "https://orcid.org/0000-0002-1814-3515"
doi: "10.1287/ijoc.2023.0281"
journal: "INFORMS Journal on Computing"
title: "Multiobjective Linear Ensembles for Robust and Sparse Training of Few-Bit Neural Networks"
year: 2024

GitHub Events

Total
Last Year