Awesome
Reconstructive Neuron Pruning for Backdoor Defense
Code for ICML 2023 Paper "Reconstructive Neuron Pruning for Backdoor Defense"
Quick Start: RNP against BadNets Attack
By default, we only use 500 defense data randomly sampled from the training set to perform the unlearn-recover process and optimize the pruning mask. To check the performance of RNP on a Badnets ResNet-18 network (i.e. 10% poisoning rata with ResNet-18 on CIFAR-10), you can directly run the command like:
python main.py
Experimental Results on BadNets Attack
[2023/07/09 22:21:00] - Namespace(alpha=0.2, arch='resnet18', backdoor_model_path='weights/ResNet18-ResNet-BadNets-target0-portion0.1-epoch80.tar', batch_size=128, clean_threshold=0.2, cuda=1, dataset='CIFAR10', log_root='logs/', mask_file=None, momentum=0.9, num_class=10, output_weight='weights/', pruning_by='threshold', pruning_max=0.9, pruning_step=0.05, ratio=0.01, recovering_epochs=20, recovering_lr=0.2, save_every=5, schedule=[10, 20], target_label=0, target_type='all2one', trig_h=3, trig_w=3, trigger_type='gridTrigger', unlearned_model_path=None, unlearning_epochs=20, unlearning_lr=0.01, weight_decay=0.0005)
[2023/07/09 22:21:00] - ----------- Data Initialization --------------
[2023/07/09 22:21:03] - ----------- Backdoor Model Initialization --------------
[2023/07/09 22:21:04] - Epoch lr Time TrainLoss TrainACC PoisonLoss PoisonACC CleanLoss CleanACC
[2023/07/09 22:21:04] - ----------- Model Unlearning --------------
[2023/07/09 22:21:15] - 0 0.010 11.2 0.1004 0.9780 0.0000 1.0000 0.2185 0.9342
[2023/07/09 22:21:26] - 1 0.010 10.8 0.1213 0.9760 0.0000 1.0000 0.2253 0.9317
[2023/07/09 22:21:37] - 2 0.010 10.8 0.1400 0.9740 0.0000 1.0000 0.2349 0.9304
[2023/07/09 22:21:48] - 3 0.010 10.8 0.1535 0.9720 0.0000 1.0000 0.2513 0.9266
[2023/07/09 22:21:59] - 4 0.010 10.9 0.2078 0.9640 0.0000 1.0000 0.2770 0.9214
[2023/07/09 22:22:10] - 5 0.010 11.0 0.2614 0.9500 0.0000 1.0000 0.3144 0.9141
[2023/07/09 22:22:21] - 6 0.010 10.8 0.3711 0.9220 0.0000 1.0000 0.3847 0.8991
[2023/07/09 22:22:31] - 7 0.010 10.8 0.4538 0.8700 0.0000 1.0000 0.5276 0.8669
[2023/07/09 22:22:42] - 8 0.010 10.8 0.7916 0.6700 0.0000 1.0000 0.9439 0.7586
[2023/07/09 22:22:53] - 9 0.010 10.8 1.4771 0.4540 0.0000 1.0000 2.5574 0.4016
[2023/07/09 22:23:04] - 10 0.001 10.8 3.1028 0.2920 0.0000 1.0000 2.5620 0.3949
[2023/07/09 22:23:15] - 11 0.001 10.8 4.0416 0.2360 0.0000 1.0000 2.8507 0.3729
[2023/07/09 22:23:25] - 12 0.001 10.8 4.8811 0.1980 0.0000 1.0000 3.2337 0.3368
[2023/07/09 22:23:25] - ----------- Model Recovering --------------
[2023/07/09 22:23:26] - Epoch lr Time TrainLoss TrainACC PoisonLoss PoisonACC CleanLoss CleanACC
[2023/07/09 22:23:37] - 1 0.200 11.0 1.0719 0.1980 0.0000 1.0000 2.0311 0.4370
[2023/07/09 22:23:48] - 2 0.200 11.0 0.9892 0.1920 0.0000 1.0000 1.3383 0.5432
[2023/07/09 22:23:59] - 3 0.200 11.0 0.7809 0.2320 0.0018 1.0000 1.1726 0.6138
[2023/07/09 22:24:10] - 4 0.200 11.0 0.4892 0.2500 0.3161 0.8770 1.0972 0.6552
[2023/07/09 22:24:21] - 5 0.200 11.1 0.4130 0.2860 0.6548 0.6051 1.0409 0.6662
[2023/07/09 22:24:32] - 6 0.200 11.0 0.3691 0.3060 0.7871 0.5084 0.9843 0.6822
[2023/07/09 22:24:43] - 7 0.200 11.0 0.3262 0.3460 0.8479 0.4791 0.9089 0.7053
[2023/07/09 22:24:54] - 8 0.200 11.0 0.2963 0.3760 0.8369 0.4904 0.8691 0.7157
[2023/07/09 22:25:05] - 9 0.200 11.0 0.2777 0.3900 0.8192 0.5090 0.8226 0.7324
[2023/07/09 22:25:16] - 10 0.200 11.0 0.2485 0.4320 0.7842 0.5340 0.7765 0.7497
[2023/07/09 22:25:27] - 11 0.200 11.0 0.2337 0.4500 0.7554 0.5562 0.7283 0.7666
[2023/07/09 22:25:38] - 12 0.200 11.0 0.2140 0.4900 0.6922 0.6044 0.7022 0.7752
[2023/07/09 22:25:49] - 13 0.200 11.0 0.2043 0.5140 0.6542 0.6317 0.6718 0.7870
[2023/07/09 22:26:00] - 14 0.200 11.0 0.1807 0.5340 0.6128 0.6598 0.6517 0.7951
[2023/07/09 22:26:11] - 15 0.200 11.0 0.1724 0.5440 0.5820 0.6873 0.6342 0.8018
[2023/07/09 22:26:22] - 16 0.200 11.0 0.1729 0.5780 0.5754 0.6968 0.6084 0.8121
[2023/07/09 22:26:33] - 17 0.200 11.1 0.1532 0.6180 0.5683 0.7027 0.5930 0.8176
[2023/07/09 22:26:44] - 18 0.200 11.0 0.1476 0.6120 0.5614 0.7083 0.5766 0.8244
[2023/07/09 22:26:55] - 19 0.200 11.0 0.1510 0.6380 0.5674 0.7069 0.5601 0.8312
[2023/07/09 22:27:06] - 20 0.200 11.1 0.1439 0.6520 0.5788 0.6988 0.5417 0.8402
[2023/07/09 22:27:07] - ----------- Backdoored Model Pruning --------------
[2023/07/09 22:27:07] - Pruned Number Layer Name Neuron Idx Mask PoisonLoss PoisonACC CleanLoss CleanACC
[2023/07/09 22:27:17] - 0 None None 0.0001 1.0000 0.2157 0.9340
[2023/07/09 22:27:27] - 12.00 layer4.1.bn2 188 0.0 0.0122 0.9986 0.2104 0.9347
[2023/07/09 22:27:37] - 14.00 layer4.1.bn2 12 0.05 0.0139 0.9982 0.2092 0.9340
[2023/07/09 22:27:46] - 18.00 layer3.0.bn2 230 0.1 0.0601 0.9876 0.2065 0.9338
[2023/07/09 22:27:56] - 21.00 bn1 5 0.15000000000000002 0.9935 0.4796 0.2074 0.9340
[2023/07/09 22:28:06] - 24.00 layer4.0.bn2 82 0.2 2.8242 0.0661 0.2297 0.9280
[2023/07/09 22:28:16] - 28.00 layer3.0.bn1 106 0.25 3.2791 0.0424 0.2292 0.9270
[2023/07/09 22:28:26] - 32.00 layer4.1.bn2 152 0.30000000000000004 4.2908 0.0172 0.2295 0.9277
Links to ImageNet-12 Subset
Please download the ImageNet-12 subset with this link: (Baidu driver)[https://pan.baidu.com/share/init?surl=LjE6g1cxQ98RZHMWi0tQlA] (pwd: qetk) or (Google driver)[https://drive.google.com/file/d/1yG9ENDUbOIUKY1i5ADu4X_7Lhbvqca2w/view?usp=sharing]
Backdoor Model Weights
You can directly download the pre-trained backdoored model weights with the links below:
| Attacks | Paper Name | Baidu Weight Source (pwd: 1212) | Google Weight Source |
|---|---|---|---|
| Badnets | Badnets: Evaluating Backdooring Attacks on Deep Neural Networks | Baidu Drive | Google Drive |
| Trojan | Trojaning attack on Neural Networks | Baidu Drive | Google Drive |
| Blend | Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoning | Baidu Drive | Google Drive |
| CL | Label-Consistent Backdoor Attacks | Baidu Drive | Google Drive |
| SIG | A New Backdoor Attack in Cnns by Training Set Corruption without Label Poisoning | Baidu Drive | Google Drive |
| Dynamic | Input-Aware Dynamic Backdoor Attack | Baidu Drive | Google Drive |
| WaNet | WaNet - Imperceptible Warping-based Backdoor Attack | Baidu Drive | Google Drive |
| FC | Poison Frog! Targeted Clean-label Backdoor Attacks on Neural Networks | Baidu Drive | Google Drive |
| DFST | Deep Feature Space Trojan Attack of Neural Networks by Controlled Detoxifcation | Baidu Drive | Google Drive |
| AWP | Can Adversarial Weight Perturbations Inject Neural Backdoors | Baidu Drive | Google Drive |
| LIRA | LIRA: Learnable, Imperceptible and Robust Backdoor Attacks | Baidu Drive | Google Drive |
| A-Blend | Circumventing Backdoor Defense that are Based on Latent Separability | Baidu Drive | Google Drive |
Citation
If you use this code in your work, please cite the accompanying paper:
@inproceedings{
li2023reconstructive,
title={Reconstructive Neuron Pruning for Backdoor Defense},
author={Yige Li and Xixiang Lyu and Xingjun Ma and Nodens Koren and Lingjuan Lyu and Bo Li and Yu-Gang Jiang},
booktitle={ICML},
year={2023},
}
Acknowledgements
As this code is reproduced based on the open-sourced code Adversarial Neuron Pruning Purifies Backdoored Deep Models and Distilling Cognitive Backdoor Patterns within an Image, the authors would like to thank their contribution and help.
Backdoor-related repo:
- Dynamic Attack: https://github.com/VinAIResearch/input-aware-backdoor-attack-release
- STRIP: https://github.com/garrisongys/STRIP
- NAD: https://github.com/bboylyg/NAD
- ABL: https://github.com/bboylyg/ABL
- Frequency: https://github.com/YiZeng623/frequency-backdoor
- NC: https://github.com/VinAIResearch/input-aware-backdoor-attack-release/tree/master/defenses/neural_cleanse
- BackdoorBox: https://github.com/THUYimingLi/BackdoorBox
- BackdoorBench:https://github.com/SCLBD/BackdoorBench