8000 Commits · 0111zizi/privacy · GitHub

More Web Proxy on the site http://driver.im/

Commits

Commits on Aug 10, 2023

Fixes comments and membership scores for thresholds attack.

shs037
authored and
tensorflower-gardener
committed

Commits on Aug 4, 2023

Merge pull request #484 from ethz-spylab/master
shs037
authored

Commits on Jul 25, 2023

Calls epsilon computation in MIA.

shs037
authored and
tensorflower-gardener
committed

Commits on Apr 29, 2023

Test DPModel in distributed training.

shs037
authored and
tensorflower-gardener
committed

Commits on Apr 13, 2023

Supports slicing for multi-label data.

shs037
authored and
tensorflower-gardener
committed

Commits on Apr 6, 2023

Skips adding noise when noise_multiplier is 0 for fast clipping.

shs037
authored and
tensorflower-gardener
committed

Commits on Apr 5, 2023

Skips noise addition when noise_multiplier is 0. Fix a typo.

shs037
authored and
tensorflower-gardener
committed

Commits on Mar 28, 2023

Sets training set as positive class for sklearn.metrics.roc_curve.

shs037
authored and
tensorflower-gardener
committed

Commits on Feb 27, 2023

Passes number of microbatches to DP model.

shs037
authored and
tensorflower-gardener
committed

Commits on Feb 25, 2023

Adds __init__.py for fast_gradient_clipping.

shs037
authored and
tensorflower-gardener
committed

Commits on Feb 16, 2023

Catches when data is not sufficient for StratifiedKFold split.

shs037
authored and
tensorflower-gardener
committed
Returns None for getting max results when results are empty.

shs037
authored and
tensorflower-gardener
committed

Commits on Feb 13, 2023

Fix a bug in get_flattened_attack_metrics that types, slices, metrics do not

shs037
authored and
tensorflower-gardener
committed

Commits on Nov 8, 2022

Allows slicing by custom indices.

shs037
authored and
tensorflower-gardener
committed

Commits on Oct 26, 2022

Adds a utility function for formating list into string.

shs037
authored and
tensorflower-gardener
committed
Add functions to derive epsilon lower bounds.

shs037
authored and
tensorflower-gardener
committed

Commits on Sep 13, 2022

Allow squared loss to take in labels and predictions of the same number of elements but different shapes.

shs037
authored and
tensorflower-gardener
committed

Commits on Aug 22, 2022

Add more documentation for `gradient_accumulation_steps` in keras optimizer.

shs037
authored and
tensorflower-gardener
committed

Commits on Aug 15, 2022

Add logging for secret sharer exposure computation.

shs037
authored and
tensorflower-gardener
committed

Commits on Aug 4, 2022

Allow specifying loss function with string.

shs037
authored and
tensorflower-gardener
committed

Commits on Jul 26, 2022

Refactor: move loss computation utilities under `privacy_tests`.

shs037
authored and
tensorflower-gardener
committed

Commits on Jul 25, 2022

Minor fix to tutorials.

shs037
authored and
tensorflower-gardener
committed

Commits on Jul 14, 2022

Add assert that the training is private for TF1 vectorized optimizer.

shs037
authored and
tensorflower-gardener
committed

Commits on Mar 3, 2022

In binary log loss for membership inference attack, allow prediction to have shape (n, 1).

shs037
authored and
tensorflower-gardener
committed

Commits on Mar 1, 2022

Adds per-example membership scores to trained attackers.

shs037
authored and
tensorflower-gardener
committed

Commits on Feb 22, 2022

Allow customized loss functions for membership inference attack.

shs037
authored and
tensorflower-gardener
committed
Fix MIA readme: labels are not needed in basic usage.

shs037
authored and
tensorflower-gardener
committed

Commits on Feb 21, 2022

Commits on Feb 16, 2022

Add links to secret sharer colab.

shs037
authored and
michaelreneer
committed

Commits on Feb 12, 2022

Functions for advanced MIAs.

shs037
authored and
tensorflower-gardener
committed

0