Privacy-Preserving Machine Learning: Threats and Solutions

Al-Rubaie, Mohammad; Chang, J. Morris

doi:10.1109/msec.2018.2888775

Cited by 281 publications

(155 citation statements)

References 24 publications

Supporting

Mentioning

153

Contrasting

Unclassified

Order By: Relevance

“…In PPDP, the explicit identifiers and non-sensitive attributes are removed before the collected data is shared with third parties, while quasi-identifiers (QIs) are generalized or suppressed, and sensitive attributes (SA) are retained as it is for the analytical purposes. In general, those who have collected and mined this data are reluctant to share this data in any way, shape, or form because attackers often posses strong background knowledge or have access to the excessive amount of auxiliary information necessary to breach safeguards used to protect user's privacy [6].…”

Section: Introductionmentioning

confidence: 99%

Attribute susceptibility and entropy based data anonymization to improve users community privacy and utility in publishing data

Majeed

Lee

2020

Appl Intell

View full text Add to dashboard Cite

User attributes affect community (i.e., a group of people with some common properties/attributes) privacy in users' data publishing because some attributes may expose multiple users' identities and their associated sensitive information during published data analysis. User attributes such as gender, age, and race, may allow an adversary to form users' communities based on their values, and launch sensitive information inference attack subsequently. As a result, explicit disclosure of private information of a specific users' community can occur from the privacy preserved published data. Each item of user attributes impacts users' community privacy differently, and some types of attributes are highly susceptible. More susceptible types of attributes enable multiple users' unique identifications and sensitive information inferences more easily, and their presence in published data increases users' community privacy risks. Most of the existing privacy models ignore the impact of susceptible attributes on user's community privacy and they mainly focus on preserving the individual privacy in the released data. This paper presents a novel data anonymization algorithm that significantly improves users' community privacy without sacrificing the guarantees on anonymous data utility in publishing data. The proposed algorithm quantifies the susceptibility of each attribute present in user's dataset to effectively preserve users' community privacy. Data generalization is performed adaptively by considering both user attributes' susceptibility and entropy simultaneously. The proposed algorithm controls overgeneralization of the data to enhance anonymous data utility for the legitimate information consumers. Due to the widespread applications of social networks (SNs), we focused on the SN users' community privacy preserved and utility enhanced anonymous data publishing. The simulation results obtained from extensive experiments, and comparisons with the existing algorithms show the effectiveness of the proposed algorithm and verify the aforementioned claims.

show abstract

Section: Introductionmentioning

confidence: 99%

Attribute susceptibility and entropy based data anonymization to improve users community privacy and utility in publishing data

Majeed

Lee

2020

Appl Intell

View full text Add to dashboard Cite

show abstract

“…These data generators could be very useful to test potential future DR policies, before deploying such solutions in consumer households. Another topic of interest is the extension of the proposed model to consider privacy of the smart meter measurements and where recent research in privacy-preserving machine learning is a promising approach [49].…”

Section: Discussionmentioning

confidence: 99%

Simulating Tariff Impact in Electrical Energy Consumption Profiles With Conditional Variational Autoencoders

Brégère

Bessa

2020

IEEE Access

View full text Add to dashboard Cite

The implementation of efficient demand response (DR) programs for household electricity consumption would benefit from data-driven methods capable of simulating the impact of different tariffs schemes. This paper proposes a novel method based on conditional variational autoencoders (CVAE) to generate, from an electricity tariff profile combined with weather and calendar variables, daily consumption profiles of consumers segmented in different clusters. First, a large set of consumers is gathered into clusters according to their consumption behavior and price-responsiveness. The clustering method is based on a causality model that measures the effect of a specific tariff on the consumption level. Then, daily electrical energy consumption profiles are generated for each cluster with CVAE. This non-parametric approach is compared to a semi-parametric data generator based on generalized additive models. Experiments in a publicly available data set show that, the proposed method presents comparable performance to the semiparametric one when it comes to generating the average value of the original data (13% difference in root mean square error). The main contribution from this new method is the capacity to reproduce rebound and side effects in the generated consumption profiles. Indeed, the application of a special electricity tariff over a time window may also affect consumption outside this time window. Another contribution is that the proposed clustering approach is capturing the reaction to a tariff change. When compared to a clustering method with classical features (min, max and average consumption), the improvement in the Calinski-Harabasz index was 128% for consumers associated with tariff changes.

show abstract

“…Protecting sensitive personal data is a major requirement in operating ML systems. In terms of active attacks, the privacy of data subjects can be at risk in multi-party environments [13], [14]. For example, a publicly available trained model can actually leak sensitive information about the training data set to an attacker who is able to run several model executions on appropriately prepared adversarial data.…”

Section: ) Privacy-sensitivitymentioning

confidence: 99%

Towards Generating Consumer Labels for Machine Learning Models

Seifert

Scherzinger

Wiese

2019

2019 IEEE First International Conference on Cognitive Machine Intelligence (CogMI)

View full text Add to dashboard Cite

Machine learning (ML) based decision making is becoming commonplace. For persons affected by ML-based decisions, a certain level of transparency regarding the properties of the underlying ML model can be fundamental. In this vision paper, we propose to issue consumer labels for trained and published ML models. These labels primarily target machine learning lay persons, such as the operators of an ML system, the executors of decisions, and the decision subjects themselves. Provided that consumer labels comprehensively capture the characteristics of the trained ML model, consumers are enabled to recognize when human intelligence should supersede artificial intelligence. In the long run, we envision a service that generates these consumer labels (semi-)automatically. In this paper, we survey the requirements that an ML system should meet, and correspondingly, the properties that an ML consumer label could capture. We further discuss the feasibility of operationalizing and benchmarking these requirements in the automated generation of ML consumer labels.

show abstract

Privacy-Preserving Machine Learning: Threats and Solutions

Cited by 281 publications

References 24 publications

Attribute susceptibility and entropy based data anonymization to improve users community privacy and utility in publishing data

Attribute susceptibility and entropy based data anonymization to improve users community privacy and utility in publishing data

Simulating Tariff Impact in Electrical Energy Consumption Profiles With Conditional Variational Autoencoders

Towards Generating Consumer Labels for Machine Learning Models

Contact Info

Product

Resources

About