Research map / Architectures and efficiency / Pruning and efficient tuning
You are caught stealing my winning lottery ticket! Making a lottery ticket claim its ownership
Xuxi Chen, Tianlong Chen, Zhenyu Zhang, Zhangyang Wang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| embed QR code signature in the sparse mask (Scheme V2) | extreme ticket ResNet-20 · CIFAR-10 | test accuracy (%): 91.66 -> 91.37 | No clear effect |
| embed QR code signature in the sparse mask (Scheme V2) | extreme ticket ResNet-18 · CIFAR-100 | test accuracy (%): 72.59 -> 72.35 | No clear effect |
| retrain with trigger set (Scheme V3) | extreme ticket ResNet-20 · CIFAR-10 trigger set | trigger accuracy (%): 16.0 -> 91.0 | Improved |
| retrain with trigger set (Scheme V3) | extreme ticket ResNet-20 · CIFAR-100 trigger set | trigger accuracy (%): 0.0 -> 90.0 | Improved |
| retrain with trigger set (Scheme V3) | extreme ticket ResNet-20 · CIFAR-100 | test accuracy (%): 66.36 -> 65.39 | Hurt |
| fine-tune key-masked ticket on other dataset (transfer attack on V1) | ResNet-20 extreme ticket · CIFAR-100 from CIFAR-10 mask | test accuracy (%): 66.36 -> 59.80 | Hurt |
| fine-tune Scheme V3 ticket on CIFAR-100 | ResNet-20 extreme ticket · CIFAR-100 | test accuracy (%): 66.36 -> 61.59 | Hurt |
| prune Scheme V3 model by 20% | ResNet-20 with trigger set · CIFAR-100 trigger set | trigger accuracy (%): 90.0 -> 16.0 | Hurt |
| one-shot magnitude pruning vs random pruning at 5% | Scheme V2 ResNet-18 · CIFAR-10 | test accuracy (%): 91.25 vs 60.87 | Mixed |
| increase OMP pruning ratio from 5% to 50% | Scheme V2 ResNet-18 · CIFAR-100 | test accuracy (%): 72.27 -> 9.24 | Hurt |
| increase add-on attack rate from 0% to 10% | ResNet-20 ticket · CIFAR-10 | test accuracy (%): 91.53 -> 71.06 | Hurt |
| increase add-on rate to 2% | ResNet-20 mask QR code · CIFAR-10 | QR decode-ability | Hurt |
Closest papers by meaning
- Sanity Checks for Lottery Tickets: Does Your Winning Ticket Really Win the Jackpot? (NeurIPS 2021 · similarity 0.89)
- Find A Winning Sign: Sign Is All We Need to Win the Lottery (ICLR 2025 · similarity 0.89)
- Coarsening the Granularity: Towards Structurally Sparse Lottery Tickets (ICML 2022 · similarity 0.88)
- Dual Lottery Ticket Hypothesis (ICLR 2022 · similarity 0.88)
- Weisfeiler and Leman Go Gambling: Why Expressive Lottery Tickets Win (ICML 2025 · similarity 0.88)
- On the Sparsity of the Strong Lottery Ticket Hypothesis (NeurIPS 2024 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
