Deep Private-Feature Extraction

Osia, Seyed Ali; Taheri, Ali; Shamsabadi, Ali Shahin; Katevas, Kleomenis; Haddadi, Hamed; Rabiee, Hamid R.

doi:10.1109/tkde.2018.2878698

Cited by 54 publications

(31 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This should motivate future work on better defenses. For instance, techniques that learn only the features relevant to a given task [15,42,43] can potentially serve as the basis for "leastprivilege" collaboratively trained models. Further, it may be possible to detect active attacks that manipulate the model into learning extra features.…”

Section: Resultsmentioning

confidence: 99%

Exploiting Unintended Feature Leakage in Collaborative Learning

Melis

Song

Cristofaro

et al. 2019

2019 IEEE Symposium on Security and Privacy (SP)

1,039

878

View full text Add to dashboard Cite

Collaborative machine learning and related techniques such as federated learning allow multiple participants, each with his own training dataset, to build a joint model by training locally and periodically exchanging model updates. We demonstrate that these updates leak unintended information about participants' training data and develop passive and active inference attacks to exploit this leakage. First, we show that an adversarial participant can infer the presence of exact data points-for example, specific locations-in others' training data (i.e., membership inference). Then, we show how this adversary can infer properties that hold only for a subset of the training data and are independent of the properties that the joint model aims to capture. For example, he can infer when a specific person first appears in the photos used to train a binary gender classifier. We evaluate our attacks on a variety of tasks, datasets, and learning configurations, analyze their limitations, and discuss possible defenses.

show abstract

Section: Resultsmentioning

confidence: 99%

Exploiting Unintended Feature Leakage in Collaborative Learning

Melis

Song

Cristofaro

et al. 2019

2019 IEEE Symposium on Security and Privacy (SP)

1,039

878

View full text Add to dashboard Cite

show abstract

“…Recently, Sanyal et al have further decreased the computational complexity of encryptionbased methods using parallelization techniques, but their method still requires more than 100 seconds to be run on a 16-machine cluster [44]. Instead of encryption-based methods, an information theoretic approach is recently introduced in [45], where the main focus is on discarding information related to a single user-defined sensitive variable. However, the end-user may not have a complete understanding about what can be inferred from her data to define as sensitive variables.…”

Section: Learning With Privacymentioning

confidence: 99%

A Hybrid Deep Learning Architecture for Privacy-Preserving Mobile Analytics

Osia

Shamsabadi

Sajadmanesh

et al. 2020

IEEE Internet Things J.

Self Cite

149

View full text Add to dashboard Cite

Internet of Things (IoT) devices and applications are being deployed in our homes and workplaces and in our daily lives. These devices often rely on continuous data collection and machine learning models for analytics and actuations. However, this approach introduces a number of privacy and efficiency challenges, as the service operator can perform arbitrary inferences on the available data. Recently, advances in edge processing have paved the way for more efficient, and private, data processing at the source for simple tasks and lighter models, though they remain a challenge for larger, and more complicated models. In this paper, we present a hybrid approach for breaking down large, complex deep neural networks for cooperative, privacy-preserving analytics. To this end, instead of performing the whole operation on the cloud, we let an IoT device to run the initial layers of the neural network, and then send the output to the cloud to feed the remaining layers and produce the final result. We manipulate the model with Siamese fine-tuning and propose a noise addition mechanism to ensure that the output of the user's device contains no extra information except what is necessary for the main task, preventing any secondary inference on the data. We then evaluate the privacy benefits of this approach based on the information exposed to the cloud service. We also asses the local inference cost of different layers on a modern handset. Our evaluations show that by using Siamese fine-tuning and at a small processing cost, we can greatly reduce the level of unnecessary, potentially sensitive information in the personal data, and thus achieving the desired trade-off between utility, privacy and performance.

show abstract

“…Transformations can reduce the amount of sensitive information in the data by reconstruction [12] or by projecting each data sample into a lower dimensional latent representation [13,28]. The information bottleneck in the hidden layers of neural networks helps to capture the main factors of variation in the data and to identify and obscure sensitive patterns in the latent representation [13], as well as during the reconstruction from the extracted low-dimensional representation [29,30].…”

Section: Filtering and Transformations Oursmentioning

confidence: 99%

Privacy and utility preserving sensor-data transformations

Malekzadeh

Clegg

Cavallaro

et al. 2020

Pervasive and Mobile Computing

Self Cite

View full text Add to dashboard Cite

Sensitive inferences and user re-identification are major threats to privacy when raw sensor data from wearable or portable devices are shared with cloud-assisted applications. To mitigate these threats, we propose mechanisms to transform sensor data before sharing them with applications running on users' devices. These transformations aim at eliminating patterns that can be used for user reidentification or for inferring potentially sensitive activities, while introducing a minor utility loss for the target application (or task). We show that, on gesture and activity recognition tasks, we can prevent inference of potentially sensitive activities while keeping the reduction in recognition accuracy of non-sensitive activities to less than 5 percentage points. We also show that we can reduce the accuracy of user re-identification and of the potential inference of gender to the level of a random guess, while keeping the accuracy of activity recognition comparable to that obtained on the original data.

show abstract

Deep Private-Feature Extraction

Cited by 54 publications

References 57 publications

Exploiting Unintended Feature Leakage in Collaborative Learning

Exploiting Unintended Feature Leakage in Collaborative Learning

A Hybrid Deep Learning Architecture for Privacy-Preserving Mobile Analytics

Privacy and utility preserving sensor-data transformations

Contact Info

Product

Resources

About