Skip to content

Consistent, fixed-width printing from adversarial crypto example. #2603

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 2 commits into from
Nov 11, 2017
Merged

Consistent, fixed-width printing from adversarial crypto example. #2603

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
284799f
Consistent, fixed-width printing from adversarial crypto example.
charlesreid1 Oct 26, 2017
a0d771b
A few more formatting fixes.
charlesreid1 Oct 27, 2017
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 5 additions & 5 deletions research/adversarial_crypto/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,16 +16,16 @@ Cryptography"](https://arxiv.org/abs/1610.06918).
> encryption and decryption, and also how to apply these operations
> selectively in order to meet confidentiality goals.

This code allows you to train an encoder/decoder/adversary triplet
This code allows you to train encoder/decoder/adversary network triplets
and evaluate their effectiveness on randomly generated input and key
pairs.

## Prerequisites

The only software requirements for running the encoder and decoder is having
Tensorflow installed.
TensorFlow installed.

Requires Tensorflow r0.12 or later.
Requires TensorFlow r0.12 or later.

## Training and evaluating

Expand All @@ -49,8 +49,8 @@ of two. In the version in the paper, there was a nonlinear unit
after the fully-connected layer; that nonlinear has been removed
here. These changes improve the robustness of training. The
initializer for the convolution layers has switched to the
tf.contrib.layers default of xavier_initializer instead of
a simpler truncated_normal.
`tf.contrib.layers default` of `xavier_initializer` instead of
a simpler `truncated_normal`.

## Contact information

Expand Down
8 changes: 4 additions & 4 deletions research/adversarial_crypto/train_eval.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -117,7 +117,7 @@ def get_message_and_key(self):
return in_m, in_k

def model(self, collection, message, key=None):
"""The model for Alice, Bob, and Eve. If key=None, the first FC layer
"""The model for Alice, Bob, and Eve. If key=None, the first fully connected layer
takes only the message as inputs. Otherwise, it uses both the key
and the message.

Expand Down Expand Up @@ -206,7 +206,7 @@ def doeval(s, ac, n, itercount):
itercount: Iteration count label for logging.

Returns:
Bob and eve's loss, as a percent of bits incorrect.
Bob and Eve's loss, as a percent of bits incorrect.
"""

bob_loss_accum = 0
Expand All @@ -217,7 +217,7 @@ def doeval(s, ac, n, itercount):
eve_loss_accum += el
bob_loss_percent = bob_loss_accum / (n * FLAGS.batch_size)
eve_loss_percent = eve_loss_accum / (n * FLAGS.batch_size)
print('%d %.2f %.2f' % (itercount, bob_loss_percent, eve_loss_percent))
print('%10d\t%20.2f\t%20.2f'%(itercount, bob_loss_percent, eve_loss_percent))
sys.stdout.flush()
return bob_loss_percent, eve_loss_percent

Expand Down Expand Up @@ -245,7 +245,7 @@ def train_and_evaluate():
with tf.Session() as s:
s.run(init)
print('# Batch size: ', FLAGS.batch_size)
print('# Iter Bob_Recon_Error Eve_Recon_Error')
print('# %10s\t%20s\t%20s'%("Iter","Bob_Recon_Error","Eve_Recon_Error"))

if train_until_thresh(s, ac):
for _ in xrange(EVE_EXTRA_ROUNDS):
Expand Down