Personal privacy vs population privacy: learning to attack anonymization

By Graham Cormode. 2011

113 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Privacy in Pharmacogenetics: An End-to-End Case Study of Personalized Warfarin Dosing (2014, neutral): “Cormode showed that if one is allowed to pose certain queries relating sensitive attributes to quasi-identifiers, it is possible to build a differentially-private Naive Bayes classifier that accurately predicts the sensitive attribute [8].”
Differential Privacy Preservation for Deep Auto-Encoders: an Application of Human Behavior Prediction (2016, methodology): “In addition, the performance of FM, FP, and DPME improves with the dataset cardinality, which is consistent with the theoretical results in (Zhang et al. 2012; Lei 2011; Cormode 2011).”
Understanding Membership Inferences on Well-Generalized Learning Models (2018, background): “se engineering attacks [17, 36] steal model parameters and structures; adversarial learning [13,22,27,35] generates misleading examples that will be misclassiﬁed by the model; model inversion attacks [10,14] infer the features of a record based on the model’s predictions on it; membership inference ”
Membership privacy: a unifying framework for privacy definitions (2013, background): “At the same time, it has been recognized that differential privacy may be too restrictive in some settings, and there are several efforts that aim at relaxing it, including differential privacy under sampling [22] and crowd blending privacy [15].”
Minimax Filter: Learning to Preserve Privacy from Inference Attacks (2016, methodology): “Furthermore, the proposed method explicitly preserves privacy against inference attacks, which both syntactic and diﬀerentially-private methods are known to be susceptible to (Cormode, 2011).”
Model Inversion Attacks that Exploit Confidence Information and Basic Countermeasures (2015, background): “Cormode showed that if an adversary is allowed to make queries that relate sensitive attributes to quasi-identifiers, then it is possible to build a differentially-private Naive Bayes classifier that accurately predicts a sensitive attribute [7].”
Differential identifiability (2012, background): “The larger the value of ǫ, the less noise needed in the model, and the more a single individual can impact that model (thus risking disclosure of information about that individual).”
Differential Privacy and Machine Learning: a Survey and Review (2014, background): “1 [9] uses privacy to mean both population privacy and individual privacy.”
EKTELO: A Framework for Defining Differentially-Private Computations (2018, background): “ond to algorithms considered in prior work. Workload represents the 2k+1 histograms as a matrix, and uses Vector Laplacetoestimatethehistogramcounts.Thiscorrespondstoatechniqueproposedintheliterature [9]. The other baseline is Identity (Plan 1): it estimates all point queries in the contingency tabl”
The Unintended Consequences of Overfitting: Training Data Inference Attacks (2017, background): “The fact that many compelling applications of this technology involve the collection and processing of sensitive personal data has given rise to concerns about privacy [1, 2, 3, 4, 5, 6, 7, 8, 9].”
Towards Measuring Membership Privacy (2017, background): “For example, model inversion attacks [14] and naive attacks [10] infer sensitive attributes of a record based on the model’s predictions on that record; membership inference attacks [37] infer whether a record is included in the training set of the model.”
Privacy Risk in Machine Learning: Analyzing the Connection to Overfitting (2017, background): “The fact that many compelling applications of this technology involve the collection and processing of sensitive personal data has given rise to concerns about privacy [1], [2], [3], [4], [5], [6], [7], [8], [9].”
Stolen Memories: Leveraging Model Memorization for Calibrated White-Box Membership Inference (2019, background): “Many compelling applications of machine learning involve the collection and processing of sensitive personal data, giving rise to concerns about privacy [2, 4, 7, 10, 11, 26, 33, 38, 45, 46].”
Preserving differential privacy in convolutional deep belief networks (2017, unknown): “2012), DPME (Lei 2011), and filter-priority (Cormode 2011).”