Marco Stronati

Bio: Marco Stronati is an academic researcher from École Polytechnique. The author has contributed to research in topics: Differential privacy & Metric (mathematics). The author has an hindex of 8, co-authored 12 publications receiving 1844 citations. Previous affiliations of Marco Stronati include French Institute for Research in Computer Science and Automation & University of Pisa.

Membership Inference Attacks Against Machine Learning Models

Abstract: We quantitatively investigate how machine learning models leak information about the individual data records on which they were trained. We focus on the basic membership inference attack: given a data record and black-box access to a model, determine if the record was in the model's training dataset. To perform membership inference against a target model, we make adversarial use of machine learning and train our own inference model to recognize differences in the target model's predictions on the inputs that it trained on versus the inputs that it did not train on. We empirically evaluate our inference techniques on classification models trained by commercial "machine learning as a service" providers such as Google and Amazon. Using realistic datasets and classification tasks, including a hospital discharge dataset whose membership is sensitive from the privacy perspective, we show that these models can be vulnerable to membership inference attacks. We then investigate the factors that influence this leakage and evaluate mitigation strategies.

Membership Inference Attacks against Machine Learning Models

Abstract: We quantitatively investigate how machine learning models leak information about the individual data records on which they were trained. We focus on the basic membership inference attack: given a data record and black-box access to a model, determine if the record was in the model's training dataset. To perform membership inference against a target model, we make adversarial use of machine learning and train our own inference model to recognize differences in the target model's predictions on the inputs that it trained on versus the inputs that it did not train on. We empirically evaluate our inference techniques on classification models trained by commercial "machine learning as a service" providers such as Google and Amazon. Using realistic datasets and classification tasks, including a hospital discharge dataset whose membership is sensitive from the privacy perspective, we show that these models can be vulnerable to membership inference attacks. We then investigate the factors that influence this leakage and evaluate mitigation strategies.

A Predictive Differentially-Private Mechanism for Mobility Traces

Abstract: With the increasing popularity of GPS-enabled handheld devices, location based applications and services have access to accurate and real-time location information, raising serious privacy concerns for their millions of users. Trying to address these issues, the notion of geo-indistinguishability was recently introduced, adapting the well-known concept of Differential Privacy to the area of location-based systems. A Laplace-based obfuscation mechanism satisfying this privacy notion works well in the case of a sporadic use; Under repeated use, however, independently applying noise leads to a quick loss of privacy due to the correlation between the location in the trace.

Constructing elastic distinguishability metrics for location privacy

Abstract: With the increasing popularity of hand-held devices, location-based applications and services have access to accurate and real-time location information, raising serious privacy concerns for their users. The recently introduced notion of geo-indistinguishability tries to address this problem by adapting the well-known concept of differential privacy to the area of location-based systems. Although geo-indistinguishability presents various appealing aspects, it has the problem of treating space in a uniform way, imposing the addition of the same amount of noise everywhere on the map. In this paper we propose a novel elastic distinguishability metric that warps the geometrical distance, capturing the different degrees of density of each area. As a consequence, the obtained mechanism adapts the level of noise while achieving the same degree of privacy everywhere. We also show how such an elastic metric can easily incorporate the concept of a "geographic fence" that is commonly employed to protect the highly recurrent locations of a user, such as his home or work. We perform an extensive evaluation of our technique by building an elastic metric for Paris' wide metropolitan area, using semantic information from the OpenStreetMap database. We compare the resulting mechanism against the Planar Laplace mechanism satisfying standard geo-indistinguishability, using two real-world datasets from the Gowalla and Brightkite location-based social networks. The results show that the elastic mechanism adapts well to the semantics of each area, adjusting the noise as we move outside the city center, hence offering better overall privacy.

Constructing elastic distinguishability metrics for location privacy

Abstract: With the increasing popularity of hand-held devices, location-based applications and services have access to accurate and real-time location information, raising serious privacy concerns for their users. The recently introduced notion of geo-indistinguishability tries to address this problem by adapting the well-known concept of differential privacy to the area of location-based systems. Although geo-indistinguishability presents various appealing aspects, it has the problem of treating space in a uniform way, imposing the addition of the same amount of noise everywhere on the map. In this paper we propose a novel elastic distinguishability metric that warps the geometrical distance, capturing the different degrees of density of each area. As a consequence, the obtained mechanism adapts the level of noise while achieving the same degree of privacy everywhere. We also show how such an elastic metric can easily incorporate the concept of a "geographic fence" that is commonly employed to protect the highly recurrent locations of a user, such as his home or work. We perform an extensive evaluation of our technique by building an elastic metric for Paris' wide metropolitan area, using semantic information from the OpenStreetMap database. We compare the resulting mechanism against the Planar Laplace mechanism satisfying standard geo-indistinguishability, using two real-world datasets from the Gowalla and Brightkite location-based social networks. The results show that the elastic mechanism adapts well to the semantics of each area, adjusting the noise as we move outside the city center, hence offering better overall privacy.

Practical Secure Aggregation for Privacy-Preserving Machine Learning

Abstract: We design a novel, communication-efficient, failure-robust protocol for secure aggregation of high-dimensional data. Our protocol allows a server to compute the sum of large, user-held data vectors from mobile devices in a secure manner (i.e. without learning each user's individual contribution), and can be used, for example, in a federated learning setting, to aggregate user-provided model updates for a deep neural network. We prove the security of our protocol in the honest-but-curious and active adversary settings, and show that security is maintained even if an arbitrarily chosen subset of users drop out at any time. We evaluate the efficiency of our protocol and show, by complexity analysis and a concrete implementation, that its runtime and communication overhead remain low even on large data sets and client pools. For 16-bit input values, our protocol offers $1.73 x communication expansion for 210 users and 220-dimensional vectors, and 1.98 x expansion for 214 users and 224-dimensional vectors over sending data in the clear.

Opportunities and obstacles for deep learning in biology and medicine.

Abstract: Deep learning describes a class of machine learning algorithms that are capable of combining raw inputs into layers of intermediate features. These algorithms have recently shown impressive results across a variety of domains. Biology and medicine are data-rich disciplines, but the data are complex and often ill-understood. Hence, deep learning techniques may be particularly well suited to solve problems of these fields. We examine applications of deep learning to a variety of biomedical problems-patient classification, fundamental biological processes and treatment of patients-and discuss whether deep learning will be able to transform these tasks or if the biomedical sphere poses unique challenges. Following from an extensive literature review, we find that deep learning has yet to revolutionize biomedicine or definitively resolve any of the most pressing challenges in the field, but promising advances have been made on the prior state of the art. Even though improvements over previous baselines have been modest in general, the recent progress indicates that deep learning methods will provide valuable means for speeding up or aiding human investigation. Though progress has been made linking a specific neural network's prediction to input features, understanding how users should interpret these models to make testable hypotheses about the system under study remains an open challenge. Furthermore, the limited amount of labelled data for training presents problems in some domains, as do legal and privacy constraints on work with sensitive health records. Nonetheless, we foresee deep learning enabling changes at both bench and bedside with the potential to transform several areas of biology and medicine.

Adversarial Examples: Attacks and Defenses for Deep Learning

Abstract: With rapid progress and significant successes in a wide spectrum of applications, deep learning is being applied in many safety-critical environments. However, deep neural networks (DNNs) have been recently found vulnerable to well-designed input samples called adversarial examples . Adversarial perturbations are imperceptible to human but can easily fool DNNs in the testing/deploying stage. The vulnerability to adversarial examples becomes one of the major risks for applying DNNs in safety-critical environments. Therefore, attacks and defenses on adversarial examples draw great attention. In this paper, we review recent findings on adversarial examples for DNNs, summarize the methods for generating adversarial examples, and propose a taxonomy of these methods. Under the taxonomy, applications for adversarial examples are investigated. We further elaborate on countermeasures for adversarial examples. In addition, three major challenges in adversarial examples and the potential solutions are discussed.

Advances and Open Problems in Federated Learning

Abstract: Federated learning (FL) is a machine learning setting where many clients (e.g. mobile devices or whole organizations) collaboratively train a model under the orchestration of a central server (e.g. service provider), while keeping the training data decentralized. FL embodies the principles of focused data collection and minimization, and can mitigate many of the systemic privacy risks and costs resulting from traditional, centralized machine learning and data science approaches. Motivated by the explosive growth in FL research, this paper discusses recent advances and presents an extensive collection of open problems and challenges.

Exploiting Unintended Feature Leakage in Collaborative Learning

Abstract: 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.