ruby/tensorflow_privacy
1
0
Fork 0
forked from 626_privacy/tensorflow_privacy
Code Pull requests Activity

Simple ROC curve plotting for membership inference attack results.

Browse source 
PiperOrigin-RevId: 325982344
This commit is contained in:
A. Unique TensorFlower 2020-08-11 02:28:48 -07:00
parent 99afaed68e
commit 680aaa4499
2 changed files with 17 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

11
tensorflow_privacy/privacy/membership_inference_attack/example.py
View file

@ -19,6 +19,8 @@ This is using a toy model based on classifying four spacial clusters of data.
"""
import os
import tempfile
import matplotlib.pyplot as plt
import numpy as np
from tensorflow import keras
from tensorflow.keras import layers
@ -28,6 +30,7 @@ from tensorflow_privacy.privacy.membership_inference_attack.data_structures impo
from tensorflow_privacy.privacy.membership_inference_attack.data_structures import AttackResults
from tensorflow_privacy.privacy.membership_inference_attack.data_structures import AttackType
from tensorflow_privacy.privacy.membership_inference_attack.data_structures import SlicingSpec
import tensorflow_privacy.privacy.membership_inference_attack.plotting as plotting
def generate_random_cluster(center, scale, num_points):
@ -147,3 +150,11 @@ print(attack_results.summary(by_slices=False))
print("Summary by slices: \n")
print(attack_results.summary(by_slices=True))
# Example of ROC curve plotting.
figure = plotting.plot_roc_curve(
attack_results.single_attack_results[0].roc_curve)
plt.show()
# For saving a figure into a file:
# plotting.save_plot(figure, <file_path>)

6
tensorflow_privacy/privacy/membership_inference_attack/plotting.py
View file

@ -78,3 +78,9 @@ def plot_histograms(train: Iterable[float],
plt.ylabel('normalized counts (density)')
plt.legend()
return fig
def plot_roc_curve(roc_curve) -> plt.Figure:
"""Plot the ROC curve and the area under the curve."""
return plot_curve_with_area(
roc_curve.fpr, roc_curve.tpr, xlabel='FPR', ylabel='TPR')