Official implementation of the paper "Backdoor Attacks on Self-Supervised Learning".

Overview

SSL-Backdoor

Abstract

Large-scale unlabeled data has allowed recent progress in self-supervised learning methods that learn rich visual representations. State-of-the-art self-supervised methods for learning representations from images (MoCo and BYOL) use an inductive bias that different augmentations (e.g. random crops) of an image should produce similar embeddings. We show that such methods are vulnerable to backdoor attacks where an attacker poisons a part of the unlabeled data by adding a small trigger (known to the attacker) to the images. The model performance is good on clean test images but the attacker can manipulate the decision of the model by showing the trigger at test time. Backdoor attacks have been studied extensively in supervised learning and to the best of our knowledge, we are the first to study them for self-supervised learning. Backdoor attacks are more practical in self-supervised learning since the unlabeled data is large and as a result, an inspection of the data to avoid the presence of poisoned data is prohibitive. We show that in our targeted attack, the attacker can produce many false positives for the target category by using the trigger at test time. We also develop a knowledge distillation based defense algorithm that succeeds in neutralizing the attack. Our code is available here: https://github.com/UMBCvision/SSL-Backdoor.

Paper

Backdoor Attacks on Self-Supervised Learning

Updates

  • 04/07/2021 - Poison generation code added.
  • 04/08/2021 - MoCo v2, BYOL code added.
  • 04/14/2021 - Jigsaw, RotNet code added.

Requirements

All experiments were run using the following dependencies.

  • python=3.7
  • pytorch=1.6.0
  • torchvision=0.7.0
  • wandb=0.10.21 (for BYOL)
  • torchnet=0.0.4 (for RotNet)

Optional

  • faiss=1.6.3 (for k-NN evaluation)

Create ImageNet-100 dataset

The ImageNet-100 dataset (random 100-class subset of ImageNet), commonly used in self-supervision benchmarks, was introduced in [1].

To create ImageNet-100 from ImageNet, use the provided script.

cd scripts
python create_imagenet_subset.py --subset imagenet100_classes.txt --full_imagenet_path <path> --subset_imagenet_path <path>

Poison Generation

To generate poisoned ImageNet-100 images, create your own configuration file. Some examples, which we use for our targeted attack experiments, are in the cfg directory.

  • You can choose the poisoning to be Targeted (poison only one category) or Untargeted
  • The trigger can be text or an image (We used triggers introduced in [2]).
  • The parameters of the trigger (e.g. location, size, alpha etc.) can be modified according to the experiment.
  • The poison injection rate for the training set can be modified.
  • You can choose which split to generate. "train" generates poisoned training data, "val_poisoned" poisons all the validation images for evaluation purpose. Note: The poisoned validation images are all resized and cropped to 224x224 before trigger pasting so that all poisoned images have uniform trigger size.
cd poison-generation
python generate_poison.py <configuration-file>

SSL Methods

Pytorch Custom Dataset

All images are loaded from filelists of the form given below.

<dir-name-1>/xxx.ext <target-class-index>
<dir-name-1>/xxy.ext <target-class-index>
<dir-name-1>/xxz.ext <target-class-index>

<dir-name-2>/123.ext <target-class-index>
<dir-name-2>/nsdf3.ext <target-class-index>
<dir-name-2>/asd932_.ext <target-class-index>

Evaluation

All evaluation scripts return confusion matrices for clean validation data and a csv file enumerating the TP and FP for each category.

MoCo v2 [3]

The implementation for MoCo is from https://github.com/SsnL/moco_align_uniform modified slightly to suit our experimental setup.

To train a ResNet-18 MoCo v2 model on ImageNet-100 on 2 NVIDIA GEFORCE RTX 2080 Ti GPUs:

cd moco
CUDA_VISIBLE_DEVICES=0,1 python main_moco.py \
                        -a resnet18 \
                        --lr 0.06 --batch-size 256 --multiprocessing-distributed \
                        --world-size 1 --rank 0 --aug-plus --mlp --cos --moco-align-w 0 \
                        --moco-unif-w 0 --moco-contr-w 1 --moco-contr-tau 0.2 \
                        --dist-url tcp://localhost:10005 \ 
                        --save-folder-root <path> \
                        --experiment-id <ID> <train-txt-file>

To train linear classifier on frozen MoCo v2 embeddings on ImageNet-100:

CUDA_VISIBLE_DEVICES=0 python eval_linear.py \
                        --arch moco_resnet18 \
                        --weights <SSL-model-checkpoint-path>\
                        --train_file <path> \
                        --val_file <path>

We use the linear classifier normalization from CompRess: Self-Supervised Learning by Compressing Representations which says "To reduce the computational overhead of tuning the hyperparameters per experiment, we standardize the Linear evaluation as following. We first normalize the features by L2 norm, then shift and scale each dimension to have zero mean and unit variance."

To evaluate linear classifier on clean and poisoned validation set: (This script loads the cached mean and variance from previous step.)

CUDA_VISIBLE_DEVICES=0 python eval_linear.py \
                        --arch moco_resnet18 \
                        --weights <SSL-model-checkpoint-path> \
                        --val_file <path> \
                        --val_poisoned_file <path> \
                        --resume <linear-classifier-checkpoint> \
                        --evaluate --eval_data <evaluation-ID> \
                        --load_cache

To run k-NN evaluation of frozen MoCo v2 embeddings on ImageNet-100 (faiss library needed):

CUDA_VISIBLE_DEVICES=0 python eval_knn.py \
                        -a moco_resnet18 \
                        --weights <SSL-model-checkpoint-path> \
                        --train_file <path> \
                        --val_file <path> \
                        --val_poisoned_file <path> \
                        --eval_data <evaluation-ID>

BYOL [4]

The implementation for BYOL is from https://github.com/htdt/self-supervised modified slightly to suit our experimental setup.

To train a ResNet-18 BYOL model on ImageNet-100 on 4 NVIDIA GEFORCE RTX 2080 Ti GPUs: (This scripts monitors the k-NN accuracy on clean ImageNet-100 dataset at regular intervals.)

cd byol
CUDA_VISIBLE_DEVICES=0,1,2,3 python -m train \
                                    --exp_id <ID> \
                                    --dataset imagenet --lr 2e-3 --emb 128 --method byol \
                                    --arch resnet18 --epoch 200 \
                                    --train_file_path <path> \
                                    --train_clean_file_path <path> 
                                    --val_file_path <path>
                                    --save_folder_root <path>

To train linear classifier on frozen BYOL embeddings on ImageNet-100:

CUDA_VISIBLE_DEVICES=0 python -m test --dataset imagenet \
                            --train_clean_file_path <path> \
                            --val_file_path <path> \
                            --emb 128 --method byol --arch resnet18 \
                            --fname <SSL-model-checkpoint-path>

To evaluate linear classifier on clean and poisoned validation set:

CUDA_VISIBLE_DEVICES=0 python -m test --dataset imagenet \
                            --val_file_path <path> \
                            --val_poisoned_file_path <path> \
                            --emb 128 --method byol --arch resnet18 \
                            --fname <SSL-model-checkpoint-path> \
                            --clf_chkpt <linear-classifier-checkpoint-path> \
                            --eval_data <evaluation-ID> --evaluate

Jigsaw [5]

The implementation for Jigsaw is our own Pytorch reimplementation based on the authors’ Caffe code https://github.com/MehdiNoroozi/JigsawPuzzleSolver modified slightly to suit our experimental setup. There might be some legacy Pytorch code, but that doesn't affect the correctness of training or evaluation. If you are looking for a recent Pytorch implementation of Jigsaw, https://github.com/facebookresearch/vissl is a good place to start.

To train a ResNet-18 Jigsaw model on ImageNet-100 on 1 NVIDIA GEFORCE RTX 2080 Ti GPU: (The code doesn't support Pytorch distributed training.)

cd jigsaw
CUDA_VISIBLE_DEVICES=0 python train_jigsaw.py \
                                --train_file <path> \
                                --val_file <path> \
                                --save <path>

To train linear classifier on frozen Jigsaw embeddings on ImageNet-100:

CUDA_VISIBLE_DEVICES=0 python eval_conv_linear.py \
                        -a resnet18 --train_file <path> \
                        --val_file <path> \
                        --save <path> \
                        --weights <SSL-model-checkpoint-path>

To evaluate linear classifier on clean and poisoned validation set:

CUDA_VISIBLE_DEVICES=0 python eval_conv_linear.py -a resnet18 \
                            --val_file <path> \
                            --val_poisoned_file <path> \
                            --weights <SSL-model-checkpoint-path> \
                            --resume <linear-classifier-checkpoint-path> \
                            --evaluate --eval_data <evaluation-ID>

RotNet [6]

The implementation for RotNet is from https://github.com/gidariss/FeatureLearningRotNet modified slightly to suit our experimental setup. There might be some legacy Pytorch code, but that doesn't affect the correctness of training or evaluation. If you are looking for a recent Pytorch implementation of RotNet, https://github.com/facebookresearch/vissl is a good place to start.

To train a ResNet-18 Jigsaw model on ImageNet-100 on 1 NVIDIA TITAN RTX GPU: (The code doesn't support Pytorch distributed training. Choose the experiment ID config file as required.)

cd rotnet
CUDA_VISIBLE_DEVICES=0 python main.py --exp <ImageNet100_RotNet_*> --save_folder <path>

To train linear classifier on frozen RotNet embeddings on ImageNet-100:

CUDA_VISIBLE_DEVICES=0 python main.py --exp <ImageNet100_LinearClassifiers_*> --save_folder <path>

To evaluate linear classifier on clean and poisoned validation set:

CUDA_VISIBLE_DEVICES=0 python main.py --exp <ImageNet100_LinearClassifiers_*> \
                            --save_folder <path> \
                            --evaluate --checkpoint=<epoch_num> --eval_data <evaluation-ID>

Acknowledgement

This material is based upon work partially supported by the United States Air Force under Contract No. FA8750‐19‐C‐0098, funding from SAP SE, NSF grant 1845216, and also financial assistance award number 60NANB18D279 from U.S. Department of Commerce, National Institute of Standards and Technology. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the United States Air Force, DARPA, or other funding agencies.

References

[1] Yonglong Tian, Dilip Krishnan, and Phillip Isola. Contrastive multiview coding. arXiv preprint arXiv:1906.05849,2019.

[2] Aniruddha Saha, Akshayvarun Subramanya, and Hamed Pirsiavash. Hidden trigger backdoor attacks. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pages 11957–11965, 2020.

[3] Chen, Xinlei, et al. "Improved baselines with momentum contrastive learning." arXiv preprint arXiv:2003.04297 (2020).

[4] Jean-Bastien Grill, Florian Strub, Florent Altch́e, and et al. Bootstrap your own latent - a new approach to self-supervised learning. In Advances in Neural Information Processing Systems, volume 33, pages 21271–21284, 2020.

[5] Noroozi, Mehdi, and Paolo Favaro. "Unsupervised learning of visual representations by solving jigsaw puzzles." European conference on computer vision. Springer, Cham, 2016.

[6] Spyros Gidaris, Praveer Singh, and Nikos Komodakis. Unsupervised representation learning by predicting image rotations. In International Conference on Learning Representations, 2018.

Citation

If you find our paper, code or models useful, please cite us using

@article{saha2021backdoor,
  title={Backdoor Attacks on Self-Supervised Learning},
  author={Saha, Aniruddha and Tejankar, Ajinkya and Koohpayegani, Soroush Abbasi and Pirsiavash, Hamed},
  journal={arXiv preprint arXiv:2105.10123},
  year={2021}
}

Questions/Issues

Please create an issue on the Github Repo directly or contact [email protected] for any questions about the code.

Owner
UMBC Vision
The Computer Vision Lab at the University of Maryland, Baltimore County (UMBC)
UMBC Vision
Web-eyes - OSINT tools for website research

WEB-EYES V1.0 web-eyes: OSINT tools for website research, 14 research methods ar

8 Nov 10, 2022
A python script to decrypt media files encrypted using the Android application 'Decrypting 'LOCKED Secret Calculator Vault''. Will identify PIN / pattern.

A python script to decrypt media files encrypted using the Android application 'Decrypting 'LOCKED Secret Calculator Vault''. Will identify PIN / pattern.

3 Sep 26, 2022
edgedressing leverages a Windows "feature" in order to force a target's Edge browser to open. This browser is then directed to a URL of choice.

edgedressing One day while experimenting with airpwn-ng, I noticed unexpected GET requests on the target node. The node in question happened to be a W

stryngs 43 Dec 23, 2022
FOSSLight Scanner performs open source analysis after downloading the source by passing a link that can be cloned by wget or git.

FOSSLight Scanner Analyze at once for Open Source Compliance. FOSSLight Scanner performs open source analysis after downloading the source by passing

FOSSLight 8 Nov 03, 2022
The Modern Hash Identification System

🔗 Don't know what type of hash it is? Name That Hash will name that hash type! 🤖 Identify MD5, SHA256 and 3000+ other hashes ☄ Comes with a neat web app 🔥

1.2k Dec 28, 2022
Use FOFA automatic vulnerability scanning tool

AutoSRC Use FOFA automatic vulnerability scanning tool Usage python3 autosrc.py -e FOFA EMAIL -k TOKEN Screenshots License MIT Dev 6613GitHub6613

PwnWiki 48 Oct 25, 2022
Tool to decrypt iOS apps using r2frida

r2flutch Yet another tool to decrypt iOS apps using r2frida. Requirements It requires to install Frida on the Jailbroken iOS device: Jailbroken device

Murphy 146 Jan 03, 2023
Cam-Hacker: Ip Cameras hack with python

Cam-Hacker Hack Cameras Mode Of Execution: apt-get install python3 apt-get insta

Error 4 You 9 Dec 17, 2022
Simple script to have LDAP authentication in Home Assistant Docker, using NGINX's ldap-auth container

Home Assistant LDAP Auth Simple script to have LDAP authentication in Home Assistant Docker, using NGINX's ldap-auth container. Usage Deploy NGINX's l

Erik 1 Sep 21, 2022
Universal Radio Hacker: Investigate Wireless Protocols Like A Boss

The Universal Radio Hacker (URH) is a complete suite for wireless protocol investigation with native support for many common Software Defined Radios.

Dr. Johannes Pohl 9k Jan 03, 2023
Infection Monkey - An automated pentest tool

Infection Monkey Data center Security Testing Tool Welcome to the Infection Monkey! The Infection Monkey is an open source security tool for testing a

Guardicore Ltd. 6k Jan 09, 2023
Lightweight and beneficial Dependency Injection plugin for apscheduler

Implementation of dependency injection for apscheduler Prerequisites: apscheduler-di solves the problem since apscheduler doesn't support Dependency I

Glib 11 Dec 07, 2022
BETA: Layla - recon tool for bug bounty

WELCOME TO LAYLA Layla is a python script that automatically performs recon on a

Matheus Faria 68 Jan 04, 2023
A collection of write-ups and solutions for Cyber FastTrack Spring 2021.

IMPORTANT: Please contact us before you use any styling or content shown here! Cyber FastTrack Spring 2021 / National Cyber Scholarship Competition -

Alice 48 Aug 28, 2022
python driver for fingerprint machine (ZKTeco biometrics)

fpmachine python driver for fingerprint machine (ZKTeco biometrics) support until now 2 model supported and tested ZMM100_TFT and ZMM220_TFT install p

Samy Sultan 4 Oct 06, 2022
Python tool for enumerating directories and for fuzzing

Python tool for enumerating directories and for fuzzing

Gourab Roy 5 Feb 21, 2022
An easy-to-use wrapper for NTFS-3G on macOS

ezNTFS ezNTFS is an easy-to-use wrapper for NTFS-3G on macOS. ezNTFS can be used as a menu bar app, or via the CLI in the terminal. Installation To us

Matthew Go 34 Dec 01, 2022
Uma ferramenta de segurança da informação escrita em python3,capaz de dar acesso total ao computador de alguém!

shell-reverse Uma ferramenta de segurança da informação escrita em python3, capaz de dar acesso total ao computador de alguém! A cybersecurity tool wr

Marcus Vinícius Ribeiro Andrade 1 Nov 03, 2021
Script Crack Facebook Premium 🚶‍♂

premium Script Crack Facebook Premium 🚶‍♂ In Script Install Script $ pkg update && pkg upgrade $ termux-setup-storage $ pkg install python $ pkg inst

Yumasaa 2 Dec 19, 2021
Fat-Stealer is a stealer that allows you to grab the Discord token from a user and open a backdoor in his machine.

Fat-Stealer is a stealer that allows you to grab the Discord token from a user and open a backdoor in his machine.

Jet Berry's 21 Jan 01, 2023