Skip to content

Commit 510faed

Browse files
statmlbenstatmlben
committed
June30|v0.15 release
1 parent 35b134b commit 510faed

File tree

5 files changed

+133
-104
lines changed

5 files changed

+133
-104
lines changed

‎README.rst

Lines changed: 3 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -101,9 +101,10 @@ Reference
101101
primaryClass={stat.ML}
102102
}
103103
104+
104105
Notebook
105106
========
106107

107-
- **MNIST dataset**: `Jupyter Notebook <https://dnn-inference.readthedocs.io/en/latest/nb/MNIST_demo.html>`_
108+
- **MNIST dataset**: `Notebook1 <https://dnn-inference.readthedocs.io/en/latest/nb/MNIST_demo.html>`_
108109

109-
- **Boston house prices dataset**: `Jupyter Notebook <https://dnn-inference.readthedocs.io/en/latest/nb/Boston_house_prices.html>`_
110+
- **Boston house prices dataset**: `Notebook2 <https://dnn-inference.readthedocs.io/en/latest/nb/Boston_house_prices.html>`_

‎dist/dnn-inference-0.15.tar.gz

22 KB
Binary file not shown.

‎dnn_inference.egg-info/PKG-INFO

Lines changed: 114 additions & 95 deletions
Original file line numberDiff line numberDiff line change
@@ -1,104 +1,123 @@
11
Metadata-Version: 2.1
22
Name: dnn-inference
3-
Version: 0.14
3+
Version: 0.15
44
Summary: Dnn-Inference is a Python module for hypothesis testing based on deep neural networks.
55
Home-page: https://github.com/statmlben/dnn-inference
66
Author: Ben Dai
77
Author-email: bendai@cuhk.edu.hk
88
License: MIT
9-
Description: .. -*- mode: rst -*-
10-
11-
|PyPi|_ |Keras|_ |MIT| |Python3| |tensorflow|_ |downloads|_ |downloads_month|_
12-
13-
.. |dAI| image:: https://img.shields.io/badge/Powered%20by-cuhk%40dAI-purple.svg
14-
.. _dAI: https://www.bendai.org
15-
16-
.. |PyPi| image:: https://badge.fury.io/py/dnn-inference.svg
17-
.. _PyPi: https://badge.fury.io/py/dnn-inference
18-
19-
.. |Keras| image:: https://img.shields.io/badge/keras-tf.keras-red.svg
20-
.. _Keras: https://keras.io/
21-
22-
.. |MIT| image:: https://img.shields.io/pypi/l/varsvm.svg
23-
24-
.. |Python3| image:: https://img.shields.io/badge/python-3-green.svg
25-
26-
.. |tensorflow| image:: https://img.shields.io/badge/keras-tensorflow-blue.svg
27-
.. _tensorflow: https://www.tensorflow.org/
28-
29-
.. |downloads| image:: https://pepy.tech/badge/dnn-inference
30-
.. _downloads: https://pepy.tech/project/dnn-inference
31-
.. |downloads_month| image:: https://pepy.tech/badge/dnn-inference/month
32-
.. _downloads_month: https://pepy.tech/project/dnn-inference
33-
34-
Dnn-Inference
35-
=============
36-
37-
.. image:: ./logo/logo_header.png
38-
:align: center
39-
:width: 800
40-
41-
Dnn-Inference is a Python module for hypothesis testing based on deep neural networks.
42-
43-
Website: https://dnn-inference.readthedocs.io
44-
45-
.. image:: ./logo/demo_result.png
46-
:align: center
47-
:width: 800
48-
49-
Three-lines-of-code
50-
-------------------
51-
.. figure:: ./logo/dnn_inf.gif
52-
53-
54-
Reference
55-
---------
56-
**If you use this code please star the repository and cite the following paper:**
57-
58-
.. code:: bib
59-
60-
@misc{dai2021significance,
61-
title={Significance tests of feature relevance for a blackbox learner},
62-
author={Ben Dai and Xiaotong Shen and Wei Pan},
63-
year={2021},
64-
eprint={2103.04985},
65-
archivePrefix={arXiv},
66-
primaryClass={stat.ML}
67-
}
68-
69-
Installation
70-
------------
71-
72-
Dependencies
73-
~~~~~~~~~~~~
74-
75-
Deep-Inference requires:
76-
77-
- Python
78-
- Numpy
79-
- Keras
80-
- Tensorflow>=2.0
81-
- sklearn
82-
- SciPy
83-
84-
User installation
85-
~~~~~~~~~~~~~~~~~
86-
87-
Install Deep-Inference using ``pip`` ::
88-
89-
pip install dnn-inference
90-
91-
or ::
92-
93-
pip install git+https://github.com/statmlben/dnn-inference.git
94-
95-
Source code
96-
~~~~~~~~~~~
97-
98-
You can check the latest sources with the command::
99-
100-
git clone https://github.com/statmlben/dnn-inference.git
101-
102-
1039
Platform: UNKNOWN
10410
Description-Content-Type: text/x-rst
11+
License-File: LICENSE
12+
13+
.. dnn-inference documentation master file
14+
15+
🔬 dnn-inference: significance tests of feature relevance for a black-box model
16+
===============================================================================
17+
18+
.. -*- mode: rst -*-
19+
20+
|PyPi|_ |Keras|_ |MIT|_ |Python3|_ |tensorflow|_ |downloads|_ |downloads_month|_
21+
22+
.. |PyPi| image:: https://badge.fury.io/py/dnn-inference.svg
23+
.. _PyPi: https://pypi.org/project/dnn-inference/
24+
25+
.. |Keras| image:: https://img.shields.io/badge/keras-tf.keras-red.svg
26+
.. _Keras: https://keras.io/
27+
28+
.. |MIT| image:: https://img.shields.io/pypi/l/dnn-inference.svg
29+
.. _MIT: https://opensource.org/licenses/MIT
30+
31+
.. |Python3| image:: https://img.shields.io/badge/python-3-green.svg
32+
.. _Python3: www.python.org
33+
34+
.. |tensorflow| image:: https://img.shields.io/badge/keras-tensorflow-blue.svg
35+
.. _tensorflow: https://www.tensorflow.org/
36+
37+
.. |downloads| image:: https://pepy.tech/badge/dnn-inference
38+
.. _downloads: https://pepy.tech/project/dnn-inference
39+
40+
.. |downloads_month| image:: https://pepy.tech/badge/dnn-inference/month
41+
.. _downloads_month: https://pepy.tech/project/dnn-inference
42+
43+
.. image:: ./logo/logo_header.png
44+
:width: 900
45+
46+
**dnn-inference** is a Python module for hypothesis testing based on black-box models, including **deep neural networks**.
47+
48+
- GitHub repo: `https://github.com/statmlben/dnn-inference <https://github.com/statmlben/dnn-inference>`_
49+
- Documentation: `https://dnn-inference.readthedocs.io <https://dnn-inference.readthedocs.io/en/latest/>`_
50+
- PyPi: `https://pypi.org/project/dnn-inference <https://pypi.org/project/nonlinear-causal>`_
51+
- Open Source: `MIT license <https://opensource.org/licenses/MIT>`_
52+
- Paper: `arXiv:2103.04985 <https://arxiv.org/abs/2103.04985>`_
53+
54+
55+
.. 🎯 What We Can Do
56+
.. -----------------
57+
58+
.. .. image:: ./logo/demo_result.png
59+
.. :width: 600
60+
61+
.. **dnn-inference** is able to provide an asymptotically valid `p-value` to examine if :math:`\mathcal{S}` is discriminative features to predict :math:`Y`.
62+
.. Specifically, the proposed testing is:
63+
64+
65+
.. H_0: R(f^*) - R_{\mathcal{S}}(g^*) = 0, \quad \text{versus} \quad H_a: R(f^*) - R_{\mathcal{S}}(g^*) < 0,
66+
67+
68+
.. where :math:`\mathcal{S}` is a collection of hypothesized features,
69+
.. :math:`R` and :math:`R_{\mathcal{S}}` are risk functions with/without the hypothesized features :math:`\mathbf{X}_{\mathcal{S}}`,
70+
.. and :math:`f^*` and :math:`g^*` are population minimizers on :math:`R` and :math:`R_{\mathcal{S}}` respectively.
71+
.. The proposed test just considers the difference between the best predictive scores with/without hypothesized features.
72+
.. Please check more details in our paper `arXiv:2103.04985 <https://arxiv.org/abs/2103.04985>`_.
73+
74+
.. - When `log-likelihood` is used as a loss function, then the test is equivalent to a conditional independence test: :math:`Y \perp X_{\mathcal{S}} | X_{\mathcal{S}^c}`.
75+
.. - Only `a small number of fitting` on neural networks is required, and the number can be as small as 1.
76+
.. - Asymptotically Type I error control and power consistency.
77+
78+
79+
Installation
80+
============
81+
82+
Dependencies
83+
------------
84+
85+
``dnn-inference`` requires: **Python>=3.8** + [pip libs](./requirements.txt)
86+
87+
.. code:: bash
88+
89+
pip install -r requirements.txt
90+
91+
User installation
92+
-----------------
93+
94+
Install ``dnn-inference`` using ``pip``
95+
96+
.. code:: bash
97+
98+
pip install dnn_inference
99+
pip install git+https://github.com/statmlben/dnn-inference.git
100+
101+
Reference
102+
---------
103+
**If you use this code please star the repository and cite the following paper:**
104+
105+
.. code:: bib
106+
107+
@misc{dai2021significance,
108+
title={Significance tests of feature relevance for a blackbox learner},
109+
author={Ben Dai and Xiaotong Shen and Wei Pan},
110+
year={2021},
111+
eprint={2103.04985},
112+
archivePrefix={arXiv},
113+
primaryClass={stat.ML}
114+
}
115+
116+
117+
Notebook
118+
========
119+
120+
- **MNIST dataset**: `Notebook1 <https://dnn-inference.readthedocs.io/en/latest/nb/MNIST_demo.html>`_
121+
122+
- **Boston house prices dataset**: `Notebook2 <https://dnn-inference.readthedocs.io/en/latest/nb/Boston_house_prices.html>`_
123+

‎dnn_inference.egg-info/SOURCES.txt

Lines changed: 5 additions & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,8 +1,12 @@
1+
LICENSE
12
README.rst
23
setup.cfg
34
setup.py
4-
dnn_inference/BBoxTest.py
5+
dnn_inference/BBox_adaTest.py
56
dnn_inference/__init__.py
7+
dnn_inference/base.py
8+
dnn_inference/old_funs.py
9+
dnn_inference/sig_test.py
610
dnn_inference.egg-info/PKG-INFO
711
dnn_inference.egg-info/SOURCES.txt
812
dnn_inference.egg-info/dependency_links.txt

‎dnn_inference.egg-info/requires.txt

Lines changed: 11 additions & 6 deletions
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,11 @@
1-
keras
2-
matplotlib
3-
numpy
4-
scipy
5-
sklearn
6-
tensorflow>=2.3.1
1+
emoji==1.7.0
2+
hachibee_sphinx_theme==0.2.5
3+
keras==2.9.0
4+
matplotlib==3.5.2
5+
numpy==1.23.0
6+
pandas==1.4.3
7+
scikit_learn==1.1.1
8+
scipy==1.8.1
9+
seaborn==0.11.2
10+
setuptools==59.6.0
11+
tensorflow==2.9.1

0 commit comments

Comments
 (0)