Skip to content
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

Jump to bottom

Update the codebase for gradient leakage attacks #377

Merged
merged 375 commits into from
Sep 18, 2024
Merged

Update the codebase for gradient leakage attacks #377

merged 375 commits into from
Sep 18, 2024

Conversation

silviafeiwang
Copy link
Collaborator

A general update for the code under the folder examples/gradient_leakage_attacks.

Description

  • Refactored the code
  • Fixed bugs
  • Incorporated more models such as ViTs
  • Ported malicious gradient leakage attacks such as the fishing attack
  • Updated .yml configuration examples
  • Added requirement.txt for pip packages required for gradient leakage attacks

How has this been tested?

Run experiments by cmd python dlg.py -c untrained_eval.yml while altering hyperparameters regarding datasets, models, attacks and defenses, etc.

Types of changes

  • Bug fix (non-breaking change which fixes an issue) Fixes #
  • New feature (non-breaking change which adds functionality)
  • Breaking change (fix or feature that would cause existing functionality to not work as expected)

Checklist:

  • My code has been formatted using Black and checked using PyLint.
  • My change requires a change to the documentation.
  • I have updated the documentation accordingly.

EthanHugh and others added 30 commits July 31, 2022 21:50
@EthanHugh
fixed the random seed setting in dlg_server
d902542
@EthanHugh
hardcoded a ground truth plot
576c98c
@EthanHugh
Bug fix branch (#14)
14c3e05 
* removed the hard coded gt plots for now, will fix the way it looks in another patch

* added our defense (Outpost) as an accepted defence

* Added proper handling for exceptions when loss becomes NaN
@EthanHugh
GT plot changes (#15)
2f7b49a 
* changed the gt plots to be saved as a pdf by default now. Added customisations to the plot for multiple images, with adge cases handled as well

* added the config file
@silviafeiwang
Merge branch 'main' of https://github.com/TL-System/plato
2c07959
@silviafeiwang
Update dlg with new changes in plato
a920671
@EthanHugh
Added another plot for the final result of the attack that follows th…
a0ada45 
…e same layout as the ground truth (which is customisable) (#16)
@silviafeiwang
Further update dlg with new changes in plato
16ce078
@silviafeiwang
Fix device and variable bugs
6e73ae0 
Still need to comment out mps device: "derivative for aten::mps_linear_backward is not implemented" error in `rec_loss.backward()`
@silviafeiwang
Merge branch 'main' of https://github.com/TL-System/plato
ebedb53
@silviafeiwang
Refactor DLG trainer
aac0c96 
to avoid changing the basic trainer
@silviafeiwang
Move files to utils folder and format files
40c0ae6
@silviafeiwang
Remove unnecessary changes to main branch Plato
48ff068
@silviafeiwang
Remove unnecessary changes to main branch Plato
4cf4083
@silviafeiwang
Refactor dlg config files
fa5cef9
@silviafeiwang
Improve dlg documentation
68f75a8
@silviafeiwang
Refactor the code using current Plato API
568f22e
@baochunli
Merge branch 'main' into plato-dlg
d837ad2
@baochunli
Added train_step_start() and perform_forward_and_backward_passes() to…
86ee3f2 
… trainer's API.
@baochunli
Added better function docstrings in the trainer.
12a66bb
@silviafeiwang
Fix bugs in trainer's APIs
63eff0d 
train_step_start() and perform_forward_and_backward_passes()
@silviafeiwang
Add outputs_generated() to trainer's API
2d35349 
and improve some function docstrings
@silviafeiwang
Refactor dlg trainer with the latest Plato API
ec6b702
@silviafeiwang
Rename private functions
0d71665
@baochunli
Some code cleanup in DLG trainer.
06fa0cc
@baochunli
Improved function docstrings.
287e708
@silviafeiwang
Update server.md with new APIs
69d7877
@silviafeiwang
Merge branch 'main' of https://github.com/TL-System/plato
fa1694a
@silviafeiwang
Fix a minor bug referring to config
dd71cd8
@silviafeiwang
Exps on LeNet5 and ResNet18
d402f29
silviafeiwang and others added 23 commits September 18, 2024 14:20
@silviafeiwang
Implemented binary attack under class collision
5dcf0f2
@silviafeiwang
Fixed a bug when modifying model weights in server payload.
3c6d98b
@silviafeiwang
Loaded all model states besides parameters in model modification.
e0e78e2
@silviafeiwang
Correct the start round for attacking a pretrained model
0ef38a6
@silviafeiwang
Plot images of target class in fishing attack
a06df34
@silviafeiwang
rPlot the GT of target images one by one
afa8728
@silviafeiwang
Removed the gap between subplots.
792e3c0
@silviafeiwang
Fixed a bug of incompatible devices.
1d3cd3f
@silviafeiwang
Incorporated delta update matching in fishing binary attack.
65164ad
@silviafeiwang
Fixed bugs of incompatible devices
d21bf13
@silviafeiwang
Added a random seed arg in .yml for Outpost perturbation.
1021b7e
@silviafeiwang
Incorporated ViT in the models.
65a8cc1
@silviafeiwang
Incorporated ViT into DLG.
dc39f48
@silviafeiwang
Updated configs
5f9e730
@silviafeiwang
Updated config files
88060d4
@silviafeiwang
Updated config files.
cf3fa12
@silviafeiwang
Merged config files.
3bb6656
@silviafeiwang
Fixed a minor bug in dlg server.
a693128
@silviafeiwang
Added requirement.txt used for dlg.
4e5f684
@silviafeiwang
Updated config files.
9193b29
@silviafeiwang
Update README.md
a6ae1da
@silviafeiwang
Format the code with Black.
7a01ef7
@silviafeiwang
Refactored all the code regarding PyLint.
b944697
@netlify Netlify
Copy link

netlify bot commented Sep 18, 2024 •
edited
Loading

✅ Deploy Preview for platodocs canceled.

Name Link
🔨 Latest commit b944697
🔍 Latest deploy log https://app.netlify.com/sites/platodocs/deploys/66eb4a86a8afb50008cfe071

@baochunli baochunli merged commit 9f0af62 into main Sep 18, 2024
7 checks passed
@baochunli baochunli deleted the gradient-leakage-attack branch September 18, 2024 22:08
@silviafeiwang silviafeiwang restored the gradient-leakage-attack branch September 20, 2024 19:47
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@baochunli baochunli Awaiting requested review from baochunli

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@silviafeiwang @baochunli @EthanHugh