https://github.com/cursedseraphim/nam-torch

A simple implementation of the Neural Additive Model by Agarwal et al. in PyTorch.
https://github.com/cursedseraphim/nam-torch

explainable-ai interpretable-ai interpretable-machine-learning iris-dataset machine-learning xai

Last synced: 4 months ago
JSON representation

A simple implementation of the Neural Additive Model by Agarwal et al. in PyTorch.

• Host: GitHub
• URL: https://github.com/cursedseraphim/nam-torch
• Owner: CursedSeraphim
• Created: 2022-09-30T11:03:56.000Z (over 3 years ago)
• Default Branch: master
• Last Pushed: 2023-04-10T16:37:16.000Z (almost 3 years ago)
• Last Synced: 2023-07-29T14:48:00.819Z (over 2 years ago)
• Topics: explainable-ai, interpretable-ai, interpretable-machine-learning, iris-dataset, machine-learning, xai
• Language: Jupyter Notebook
• Homepage:
• Size: 846 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Neural Additive Models for PyTorch

An implementation of a simple [Neural Additive Model](https://neural-additive-models.github.io/). ([Agarwal et al. 2020](https://arxiv.org/abs/2004.13912)) in PyTorch. It is applied to the Iris dataset for classification and feature map visualization.

![iris_feature_maps](https://user-images.githubusercontent.com/29144639/193257422-80744e25-f5cb-4824-8e46-96a27a2f5a61.png)

# TODOs

* Implement the ExU

* Implement Binary Classification model structure

* Implement Regression model structure

* Implement a more general framework / classes / functions that perform the feature map visualization on arbitrary datasets