tribhuvanesh / knockoffnets
Knockoff Nets: Stealing Functionality of Black-Box Models
☆89Updated last year
Related projects: ⓘ
- This is an implementation demo of the ICLR 2021 paper [Neural Attention Distillation: Erasing Backdoor Triggers from Deep Neural Networks…☆115Updated 2 years ago
- CVPR 2021 Official repository for the Data-Free Model Extraction paper. https://arxiv.org/abs/2011.14779☆66Updated 5 months ago
- The code is for our NeurIPS 2019 paper: https://arxiv.org/abs/1910.04749☆31Updated 4 years ago
- Fine-Pruning: Defending Against Backdooring Attacks on Deep Neural Networks (RAID 2018)☆47Updated 5 years ago
- Attacking a dog vs fish classification that uses transfer learning inceptionV3☆67Updated 6 years ago
- Prediction Poisoning: Towards Defenses Against DNN Model Stealing Attacks (ICLR '20)☆29Updated 3 years ago
- Official Implementation of ICLR 2022 paper, ``Adversarial Unlearning of Backdoors via Implicit Hypergradient''☆49Updated last year
- This is for releasing the source code of the ACSAC paper "STRIP: A Defence Against Trojan Attacks on Deep Neural Networks"☆47Updated 4 years ago
- ☆88Updated 3 years ago
- Code for the paper: Label-Only Membership Inference Attacks☆61Updated 3 years ago
- ☆76Updated 3 years ago
- ☆44Updated 3 years ago
- Code for "Diversity can be Transferred: Output Diversification for White- and Black-box Attacks"☆52Updated 3 years ago
- Input-aware Dynamic Backdoor Attack (NeurIPS 2020)☆27Updated last month
- ABS: Scanning Neural Networks for Back-doors by Artificial Brain Stimulation☆47Updated 2 years ago
- ☆48Updated 2 years ago
- Code for "On Adaptive Attacks to Adversarial Example Defenses"☆84Updated 3 years ago
- Code for "Label-Consistent Backdoor Attacks"☆48Updated 3 years ago
- Attacks Which Do Not Kill Training Make Adversarial Learning Stronger (ICML2020 Paper)☆124Updated last year
- ☆73Updated 3 years ago
- Witches' Brew: Industrial Scale Data Poisoning via Gradient Matching☆91Updated last month
- ☆31Updated 2 weeks ago
- Implementation of the Model Inversion Attack introduced with Model Inversion Attacks that Exploit Confidence Information and Basic Counte…☆79Updated last year
- Craft poisoned data using MetaPoison☆47Updated 3 years ago
- [NeurIPS 2019] This is the code repo of our novel passport-based DNN ownership verification schemes, i.e. we embed passport layer into va…☆77Updated last year
- Implementation of the paper "MAZE: Data-Free Model Stealing Attack Using Zeroth-Order Gradient Estimation".☆28Updated 2 years ago
- Target Agnostic Attack on Deep Models: Exploiting Security Vulnerabilities of Transfer Learning☆9Updated 5 years ago
- ☆60Updated 3 years ago
- A unified benchmark problem for data poisoning attacks☆148Updated 11 months ago
- ☆41Updated last year