were / cs239

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

DUTI Implementation in Python

Data set

Income data set is already included in this repository.

You could download spam data from Colab for Influence Function

Files

├── exp_income.ipynb (Experiment for fixing data of income)
├── exp_sincurve.ipynb (Experiment for fixing sin(x))
├── interactive_mnist_duti.ipynb (Interactive UI for fixing data of MNIST using traditional DUTI)
├── interactive_mnist_iduti.ipynb (Interactive UI for fixing data of MNIST using iDUTI)
├── fig_input_checking.ipynb (Plotting the figure of MNIST input checking)
├── fig_spam_input_checking.ipynb (Plotting the figure of spam input checking)
├── exp_auto_fix_spam.ipynb (Experiment for automatically fixing spam)
├── influence (Library directly use from Influence Function)
├── duti (Implemented Library of DUTI and its experiments)
│   ├── duti.py
│   ├── experiments.py
│   └── __init__.py
├── data (Supporting testset)
│   ├── adult.csv
│   └── spam
└── utils (Util functions)
    ├── experiment_mnist.py
    ├── __init__.py
    ├── load_mnist.py
    ├── load_spam.py
    └── mnist_devel.ipynb

About


Languages

Language:Python 51.0%Language:Jupyter Notebook 49.0%