We lifted 9 functions out of this paper's own repositories and ran 5 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.
| Repository | Role | Ran |
|---|---|---|
| rfww/spy-watermark | canonical | 5 of 9 |
| Function | Status | Where it lives |
|---|---|---|
| Get_loss_func | Ran | rfww/spy-watermark/loss.py pointer only (licence: NONE) · get_code("c447b70e0c6bc034") |
| Get_loss_func | Ran | rfww/spy-watermark/optimizer.py pointer only (licence: NONE) · get_code("f085d3a1f8f83a06") |
| Get_optimizers | Ran | rfww/spy-watermark/optimizer.py pointer only (licence: NONE) · get_code("186c2b5c3a2921c8") |
| PSNR | Ran | rfww/spy-watermark/utils.py pointer only (licence: NONE) · get_code("c898fb896cfeaece") |
| crop | Ran | rfww/spy-watermark/functional.py pointer only (licence: NONE) · get_code("29362ecc361d2231") |
| Get_dataloader_test | Not yet run | rfww/spy-watermark/datasets.py pointer only (licence: NONE) · get_code("f9cc801a4279902a") |
| get_1d_sincos_pos_embed_from_grid | Not yet run | rfww/spy-watermark/models/pos_embed.py pointer only (licence: NONE) · get_code("12035a2f77d8016c") |
| get_2d_sincos_pos_embed | Not yet run | rfww/spy-watermark/models/pos_embed.py pointer only (licence: NONE) · get_code("3185afc3e87293ed") |
| get_2d_sincos_pos_embed_from_grid | Not yet run | rfww/spy-watermark/models/pos_embed.py pointer only (licence: NONE) · get_code("f10004e059714d42") |
Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.
Backdoor attack aims to deceive a victim model when facing backdoor instances while maintaining its performance on benign data. Current methods use manual patterns or special perturbations as triggers, while they often overlook the robustness against data corruption, making backdoor attacks easy to defend in practice. To address this issue, we propose a novel backdoor attack method named Spy-Watermark, which remains effective when facing data collapse and backdoor defense. Therein, we introduce a learnable watermark embedded in the latent domain of images, serving as the trigger. Then, we search for a watermark that can withstand collapse during image decoding, cooperating with several anti-collapse operations to further enhance the resilience of our trigger against data corruption. Extensive experiments are conducted on CIFAR10, GTSRB, and ImageNet datasets, demonstrating that Spy-Watermark overtakes ten state-of-the-art methods in terms of robustness and stealthiness.
The same record, over MCP at https://syntology.ai/mcp:
get_harvested_code_for_paper("2401.02031")
get_code_for_paper("2401.02031")
have("2401.02031")
Connect an agent — have() is free.