his.sePublications
Change search
Link to record
Permanent link

Direct link
BETA
Senavirathne, NavodaORCID iD iconorcid.org/0000-0002-2564-0683
Publications (4 of 4) Show all publications
Senavirathne, N. (2019). Addressing the challenges of privacy preserving machine learning in the context of data anonymization. Skövde: University of Skövde
Open this publication in new window or tab >>Addressing the challenges of privacy preserving machine learning in the context of data anonymization
2019 (English)Report (Other academic)
Abstract [en]

Machine learning (ML) models trained on sensitive data pose a distinct threat to privacy with the emergence of numerous threat models exploiting their privacy vulnerabilities.Therefore, privacy preserving machine learning (PPML) has gained an increased attentionover the past couple of years. Existing PPML techniques introduced in the literatureare mainly based on differential privacy or cryptography based techniques. Respectivelythey are criticized for the poor predictive accuracy of the derived ML models and for theextensive computational cost. Moreover, they operate under the assumption that originaldata are always available for training the ML models. However, there exist scenarioswhere anonymized data are available instead of the original data. Anonymization ofsensitive data is required before publishing them in order to preserve the privacy of theunderlying data subjects. Nevertheless, there are valid organizational and legal requirementsfor data publishing. In this case, it is important to understand the impact of dataanonymization on ML in general and how this can be used as a stepping stone towardsPPML.The proposed research is aimed at understanding the opportunities and challenges forPPML in the context of data anonymization, and to address them effectively by developinga unified solution to serve the objectives of both data anonymization and PPML.

Place, publisher, year, edition, pages
Skövde: University of Skövde, 2019. p. 60
Keywords
privacy preserving machine learning, privacy preserving data publishing, data anonymization, privacy vulnerabilities in machine learning
National Category
Computer Systems
Research subject
Skövde Artificial Intelligence Lab (SAIL)
Identifiers
urn:nbn:se:his:diva-16815 (URN)
Note

Research proposal, PhD programme, University of Skövde

Available from: 2019-04-24 Created: 2019-04-24 Last updated: 2019-05-02Bibliographically approved
Senavirathne, N. & Torra, V. (2019). Integrally private model selection for decision trees. Computers & security (Print), 83, 167-181
Open this publication in new window or tab >>Integrally private model selection for decision trees
2019 (English)In: Computers & security (Print), ISSN 0167-4048, E-ISSN 1872-6208, Vol. 83, p. 167-181Article in journal (Refereed) Published
Abstract [en]

Privacy attacks targeting machine learning models are evolving. One of the primary goals of such attacks is to infer information about the training data used to construct the models. “Integral Privacy” focuses on machine learning and statistical models which explain how we can utilize intruder's uncertainty to provide a privacy guarantee against model comparison attacks. Through experimental results, we show how the distribution of models can be used to achieve integral privacy. Here, we observe two categories of machine learning models based on their frequency of occurrence in the model space. Then we explain the privacy implications of selecting each of them based on a new attack model and empirical results. Also, we provide recommendations for private model selection based on the accuracy and stability of the models along with the diversity of training data that can be used to generate the models. 

Place, publisher, year, edition, pages
Elsevier Ltd, 2019
Keywords
Data privacy, Integral privacy, Machine learning model space, Privacy models, Privacy preserving machine learning, Decision trees, Attack model, Machine learning models, Model comparison, Model Selection, Privacy Attacks, Privacy preserving, Training data, Machine learning
National Category
Computer Sciences
Research subject
Skövde Artificial Intelligence Lab (SAIL)
Identifiers
urn:nbn:se:his:diva-16682 (URN)10.1016/j.cose.2019.01.006 (DOI)000465367100013 ()2-s2.0-85062062700 (Scopus ID)
Available from: 2019-03-08 Created: 2019-03-08 Last updated: 2019-07-10Bibliographically approved
Torra, V. & Senavirathne, N. (2019). Maximal c consensus meets. Information Fusion, 51, 58-66
Open this publication in new window or tab >>Maximal c consensus meets
2019 (English)In: Information Fusion, ISSN 1566-2535, E-ISSN 1872-6305, Vol. 51, p. 58-66Article in journal (Refereed) Published
Abstract [en]

Given a set S of subsets of a reference set X, we define the problem of finding c subsets of X that maximize the size of the intersection among the included subsets. Maximizing the size of the intersection means that they are subsets of the sets in S and they are as large as possible. We can understand the result of this problem as c consensus sets of S, or c consensus representatives of S. From the perspective of lattice theory, each representative will be a meet of some sets in S. In this paper we define formally this problem, and present heuristic algorithms to solve it. We also discuss the relationship with other established problems in the literature.

Place, publisher, year, edition, pages
NETHERLANDS: Elsevier, 2019
Keywords
clustering, consensus clustering, heuristic algorithms, Maximal c consensus meets, Cluster analysis, Clustering algorithms, Lattice theory, Set theory, Reference set
National Category
Computer Sciences
Research subject
Skövde Artificial Intelligence Lab (SAIL)
Identifiers
urn:nbn:se:his:diva-16463 (URN)10.1016/j.inffus.2018.09.011 (DOI)000469155600006 ()2-s2.0-85056612105 (Scopus ID)
Available from: 2019-01-30 Created: 2019-01-30 Last updated: 2019-07-10Bibliographically approved
Senavirathne, N. & Torra, V. (2018). Approximating Robust Linear Regression With An Integral Privacy Guarantee. In: Kieran McLaughlin, Ali Ghorbani, Sakir Sezer, Rongxing Lu, Liqun Chen, Robert H. Deng, Paul Miller, Stephen Marsh, Jason Nurse (Ed.), 2018 16th Annual Conference on Privacy, Security and Trust (PST): . Paper presented at 16th Annual Conference on Privacy, Security and Trust (PST), Belfast, Northern Ireland, August 28-30, 2018 (pp. 85-94). IEEE
Open this publication in new window or tab >>Approximating Robust Linear Regression With An Integral Privacy Guarantee
2018 (English)In: 2018 16th Annual Conference on Privacy, Security and Trust (PST) / [ed] Kieran McLaughlin, Ali Ghorbani, Sakir Sezer, Rongxing Lu, Liqun Chen, Robert H. Deng, Paul Miller, Stephen Marsh, Jason Nurse, IEEE, 2018, p. 85-94Conference paper, Published paper (Refereed)
Abstract [en]

Most of the privacy-preserving techniques suffer from an inevitable utility loss due to different perturbations carried out on the input data or the models in order to gain privacy. When it comes to machine learning (ML) based prediction models, accuracy is the key criterion for model selection. Thus, an accuracy loss due to privacy implementations is undesirable. The motivation of this work, is to implement the privacy model "integral privacy" and to evaluate its eligibility as a technique for machine learning model selection while preserving model utility. In this paper, a linear regression approximation method is implemented based on integral privacy which ensures high accuracy and robustness while maintaining a degree of privacy for ML models. The proposed method uses a re-sampling based estimator to construct linear regression model which is coupled with a rounding based data discretization method to support integral privacy principles. The implementation is evaluated in comparison with differential privacy in terms of privacy, accuracy and robustness of the output ML models. In comparison, integral privacy based solution provides a better solution with respect to the above criteria.

Place, publisher, year, edition, pages
IEEE, 2018
Series
Annual Conference on Privacy Security and Trust-PST, ISSN 1712-364X
Keywords
Integral privacy, Linear regression, Privacy-preserving machine learning
National Category
Computer Sciences
Research subject
Skövde Artificial Intelligence Lab (SAIL); INF301 Data Science; INF303 Information Security
Identifiers
urn:nbn:se:his:diva-16573 (URN)10.1109/PST.2018.8514161 (DOI)000454683600008 ()2-s2.0-85063441298 (Scopus ID)978-1-5386-7494-9 (ISBN)978-1-5386-7493-2 (ISBN)
Conference
16th Annual Conference on Privacy, Security and Trust (PST), Belfast, Northern Ireland, August 28-30, 2018
Available from: 2019-01-18 Created: 2019-01-18 Last updated: 2019-07-10Bibliographically approved
Organisations
Identifiers
ORCID iD: ORCID iD iconorcid.org/0000-0002-2564-0683

Search in DiVA

Show all publications