From 8a9fec89225caa0974313950f52ef0a97f3f13b8 Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Thu, 15 Apr 2021 15:31:50 -0400 Subject: [PATCH 01/13] Add split NN inversion paper --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 690f6a9..97a698c 100644 --- a/README.md +++ b/README.md @@ -133,6 +133,7 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**KART: Privacy Leakage Framework of Language Models Pre-trained with Clinical Records**](https://arxiv.org/abs/2101.00036) (Nakamura et al., 2020) - [**Derivation of Constraints from Machine Learning Models and Applications to Security and Privacy**](https://hal.archives-ouvertes.fr/hal-03091740/) (Falaschi et al., 2021) - [**On the (In)Feasibility of Attribute Inference Attacks on Machine Learning Models**](https://arxiv.org/abs/2103.07101) (Zhao et al., 2021) +- [**Practical Defences Against Model Inversion Attacks for Split Neural Networks**](https://arxiv.org/abs/2104.05743) (Titcombe et al., 2021) ## Property inference From 68669f9091d8db62ba49028d567c34059ec9de53 Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Sat, 17 Apr 2021 07:11:05 -0400 Subject: [PATCH 02/13] Update README.md --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 97a698c..753bc97 100644 --- a/README.md +++ b/README.md @@ -187,6 +187,7 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Protecting Decision Boundary of Machine Learning Model With Differentially Private Perturbation**](https://ieeexplore.ieee.org/abstract/document/9286504) (Zheng et al., 2021) - [**Special-Purpose Model Extraction Attacks: Stealing Coarse Model with Fewer Queries**](https://ieeexplore.ieee.org/abstract/document/9343086?casa_token=Fn4CtwOZsbQAAAAA:4n3tZGcwFochwREqn4fRWcmA9YeLRxikwB1LN8t2ui1NbRPHSHjTuoqHrSfP1vxXfecw0kobBQ) (Okada et al., 2021) - [**Model Extraction and Adversarial Transferability, Your BERT is Vulnerable!**](https://arxiv.org/abs/2103.10013) (He et al., 2021) ([code](https://github.com/xlhex/extract_and_transfer)) +- [**Thief, Beware of What Get You There: Towards Understanding Model Extraction Attack**](https://arxiv.org/abs/2104.05921) (Zhang et al., 2021) # Other From c487a642d928ba0ca1ba8d17095d14a1165436d5 Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Sat, 17 Apr 2021 07:13:22 -0400 Subject: [PATCH 03/13] Update README.md --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 753bc97..45a6074 100644 --- a/README.md +++ b/README.md @@ -84,6 +84,7 @@ This repository contains a curated list of papers related to privacy attacks aga - [**Node-Level Membership Inference Attacks Against Graph Neural Networks**](https://arxiv.org/abs/2102.05429) (He et al., 2021) - [**Practical Blind Membership Inference Attack via Differential Comparisons**](https://arxiv.org/abs/2101.01341) (Hui et al., 2021) - [**ADePT: Auto-encoder based Differentially Private Text Transformation**](https://arxiv.org/abs/2102.01502) (Krishna et al., 2021) +- [**The Influence of Dropout on Membership Inference in Differentially Private Models**](https://arxiv.org/abs/2103.09008) (Galinkin, 2021) ## Reconstruction From 5ae97c33df93853d733ff5580fa15c20a112e95c Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Mon, 3 May 2021 11:54:04 -0400 Subject: [PATCH 04/13] Add privacy paper --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 45a6074..2564cce 100644 --- a/README.md +++ b/README.md @@ -203,3 +203,4 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Quantifying Privacy Leakage in Graph Embedding**](https://arxiv.org/abs/2010.00906) (Duddu et al., 2020) - [**Quantifying and Mitigating Privacy Risks of Contrastive Learning**](https://arxiv.org/abs/2102.04140) (He and Zhang, 2021) - [**Coded Machine Unlearning**](https://arxiv.org/abs/2012.15721) (Aldaghri et al., 2020) +- [**Unlearnable Examples: Making Personal Data Unexploitable**](https://arxiv.org/abs/2101.04898) (Huang et al., 2021) From 1821a1c33690315463c3c1a94e72bb1541131a4e Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Mon, 3 May 2021 19:13:20 -0400 Subject: [PATCH 05/13] add paper --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 2564cce..259ed9d 100644 --- a/README.md +++ b/README.md @@ -135,6 +135,7 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Derivation of Constraints from Machine Learning Models and Applications to Security and Privacy**](https://hal.archives-ouvertes.fr/hal-03091740/) (Falaschi et al., 2021) - [**On the (In)Feasibility of Attribute Inference Attacks on Machine Learning Models**](https://arxiv.org/abs/2103.07101) (Zhao et al., 2021) - [**Practical Defences Against Model Inversion Attacks for Split Neural Networks**](https://arxiv.org/abs/2104.05743) (Titcombe et al., 2021) +- [**R-GAP: Recursive Gradient Attack on Privacy**](https://arxiv.org/abs/2010.07733) (Zhu and Blaschko, 2021) ## Property inference From 9dc685703cd6d6c65da299299167fcc51f845652 Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Mon, 3 May 2021 19:39:55 -0400 Subject: [PATCH 06/13] Update README.md --- README.md | 2 ++ 1 file changed, 2 insertions(+) diff --git a/README.md b/README.md index 259ed9d..bfacba2 100644 --- a/README.md +++ b/README.md @@ -190,6 +190,8 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Special-Purpose Model Extraction Attacks: Stealing Coarse Model with Fewer Queries**](https://ieeexplore.ieee.org/abstract/document/9343086?casa_token=Fn4CtwOZsbQAAAAA:4n3tZGcwFochwREqn4fRWcmA9YeLRxikwB1LN8t2ui1NbRPHSHjTuoqHrSfP1vxXfecw0kobBQ) (Okada et al., 2021) - [**Model Extraction and Adversarial Transferability, Your BERT is Vulnerable!**](https://arxiv.org/abs/2103.10013) (He et al., 2021) ([code](https://github.com/xlhex/extract_and_transfer)) - [**Thief, Beware of What Get You There: Towards Understanding Model Extraction Attack**](https://arxiv.org/abs/2104.05921) (Zhang et al., 2021) +- [**Model Weight Theft With Just Noise Inputs: The Curious Case of the Petulant Attacker**](https://arxiv.org/abs/1912.08987) (Roberts et al., 2019) +- [**Protecting DNNs from Theft using an Ensemble of Diverse Models**](https://openreview.net/forum?id=LucJxySuJcE) (Kariyappa et al., 2021) # Other From 220b63e62914f0cdaa1cffb0c5f55cf79b1cb34c Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Tue, 4 May 2021 13:27:19 -0400 Subject: [PATCH 07/13] Update README.md --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index bfacba2..f808442 100644 --- a/README.md +++ b/README.md @@ -192,6 +192,7 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Thief, Beware of What Get You There: Towards Understanding Model Extraction Attack**](https://arxiv.org/abs/2104.05921) (Zhang et al., 2021) - [**Model Weight Theft With Just Noise Inputs: The Curious Case of the Petulant Attacker**](https://arxiv.org/abs/1912.08987) (Roberts et al., 2019) - [**Protecting DNNs from Theft using an Ensemble of Diverse Models**](https://openreview.net/forum?id=LucJxySuJcE) (Kariyappa et al., 2021) +- [**Information Laundering for Model Privacy**](https://arxiv.org/abs/2009.06112) (Wang et al., 2021) # Other From 38d05fc0d92565d1a107ab6c077a5c436333734c Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Tue, 4 May 2021 14:13:19 -0400 Subject: [PATCH 08/13] Update README.md --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index f808442..49e022a 100644 --- a/README.md +++ b/README.md @@ -193,6 +193,7 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Model Weight Theft With Just Noise Inputs: The Curious Case of the Petulant Attacker**](https://arxiv.org/abs/1912.08987) (Roberts et al., 2019) - [**Protecting DNNs from Theft using an Ensemble of Diverse Models**](https://openreview.net/forum?id=LucJxySuJcE) (Kariyappa et al., 2021) - [**Information Laundering for Model Privacy**](https://arxiv.org/abs/2009.06112) (Wang et al., 2021) +- [**Deep Neural Network Fingerprinting by Conferrable Adversarial Examples**](https://arxiv.org/abs/1912.00888) (Lukas et al., 2021) # Other From 69d27b10770fb05b424541f1103befb0f039ae56 Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Wed, 5 May 2021 16:53:15 -0400 Subject: [PATCH 09/13] Add property inference paper --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 49e022a..df86247 100644 --- a/README.md +++ b/README.md @@ -145,6 +145,7 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Overlearning Reveals Sensitive Attributes**](https://openreview.net/pdf?id=SJeNz04tDS) (Song C. et al., 2020) ([code](https://drive.google.com/file/d/1hu0PhN3pWXe6LobxiPFeYBm8L-vQX2zJ/view?usp=sharing)) - [**Subject Property Inference Attack in Collaborative Learning**](https://ieeexplore.ieee.org/document/9204357) (Xu and Li, 2020) - [**Property Inference From Poisoning**](https://arxiv.org/abs/2101.11073) (Chase et al., 2021) +- [**Property Inference Attacks on Convolutional Neural Networks: Influence and Implications of Target Model's Complexity**](https://arxiv.org/abs/2104.13061) (Parisot et al., 2021) ## Model extraction From 465ac0ed2478cabbfb360f578aa1674206de0d74 Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Wed, 5 May 2021 16:56:31 -0400 Subject: [PATCH 10/13] Add model inversion paper --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index df86247..42122f3 100644 --- a/README.md +++ b/README.md @@ -136,6 +136,7 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**On the (In)Feasibility of Attribute Inference Attacks on Machine Learning Models**](https://arxiv.org/abs/2103.07101) (Zhao et al., 2021) - [**Practical Defences Against Model Inversion Attacks for Split Neural Networks**](https://arxiv.org/abs/2104.05743) (Titcombe et al., 2021) - [**R-GAP: Recursive Gradient Attack on Privacy**](https://arxiv.org/abs/2010.07733) (Zhu and Blaschko, 2021) +- [**Exploiting Explanations for Model Inversion Attacks**](https://arxiv.org/abs/2104.12669) (Zhao et al., 2021) ## Property inference From 1b9dec6a4972169b011e3be3cde6c28f68251ab7 Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Wed, 5 May 2021 22:48:27 -0400 Subject: [PATCH 11/13] Update README.md --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 42122f3..70a707a 100644 --- a/README.md +++ b/README.md @@ -196,6 +196,7 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Protecting DNNs from Theft using an Ensemble of Diverse Models**](https://openreview.net/forum?id=LucJxySuJcE) (Kariyappa et al., 2021) - [**Information Laundering for Model Privacy**](https://arxiv.org/abs/2009.06112) (Wang et al., 2021) - [**Deep Neural Network Fingerprinting by Conferrable Adversarial Examples**](https://arxiv.org/abs/1912.00888) (Lukas et al., 2021) +- [**BODAME: Bilevel Optimization for Defense Against Model Extraction**](https://arxiv.org/abs/2103.06797) (Mori et al., 2021) # Other From cc0b7cdca31180ff50a4a96807f239e1895699fd Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Thu, 6 May 2021 01:44:19 -0400 Subject: [PATCH 12/13] Add model extraction paper --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index 70a707a..a9cc269 100644 --- a/README.md +++ b/README.md @@ -197,6 +197,7 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Information Laundering for Model Privacy**](https://arxiv.org/abs/2009.06112) (Wang et al., 2021) - [**Deep Neural Network Fingerprinting by Conferrable Adversarial Examples**](https://arxiv.org/abs/1912.00888) (Lukas et al., 2021) - [**BODAME: Bilevel Optimization for Defense Against Model Extraction**](https://arxiv.org/abs/2103.06797) (Mori et al., 2021) +- [**Dataset Inference: Ownership Resolution in Machine Learning**](https://openreview.net/forum?id=hvdKKV2yt7T) (Maini et al., 2021) # Other From 3eba3c558d29d4c0c0933505d338b692d93d1c17 Mon Sep 17 00:00:00 2001 From: Suha Sabi Hussain Date: Tue, 11 May 2021 17:22:56 -0400 Subject: [PATCH 13/13] Update README.md --- README.md | 1 + 1 file changed, 1 insertion(+) diff --git a/README.md b/README.md index a9cc269..2207db0 100644 --- a/README.md +++ b/README.md @@ -213,3 +213,4 @@ Reconstruction attacks cover also attacks known as *model inversion* and *attrib - [**Quantifying and Mitigating Privacy Risks of Contrastive Learning**](https://arxiv.org/abs/2102.04140) (He and Zhang, 2021) - [**Coded Machine Unlearning**](https://arxiv.org/abs/2012.15721) (Aldaghri et al., 2020) - [**Unlearnable Examples: Making Personal Data Unexploitable**](https://arxiv.org/abs/2101.04898) (Huang et al., 2021) +- [**Measuring Data Leakage in Machine-Learning Models with Fisher Information**](https://arxiv.org/abs/2102.11673) (Hannun et al., 2021)