Differential privacy for AI & ML models

Although the implementation of privacy regulations and laws is not a recent development, new laws are being implemented to the challenges and characteristics of the modern information era.

Benodigde interesse(s)

  • Software Architecture
  • Software Development Methodologies
  • Artificial Intelligence

Wat krijg je

  • Een uitdagende afstudeeropdracht
  • € 1000 vergoeding of € 500 + leaseauto
  • Professionele begeleiding
  • Trainingen gericht op jouw afstudeertraject
  • Sparren met andere afstudeerders en collega’s
  • Twee vakantiedagen per maand

Wat je gaat doen

65% Afstudeeronderzoek

An example of a relatively new law is the General Data Protection Regulation (GDPR). This regulation creates more conditions for data processors, more rights for the individuals whose data is being processed, and larger fines for violations. Therefor the need for data anonymising is increasing. A way to anonymise data is by implementing differential privacy. Differential privacy is a statistical technique that aims to provide means to maximise the accuracy of queries from statistical databases while measuring (and, thereby, hopefully minimising) the privacy impact on individuals whose information is in the database. The implementation of differential privacy may impact the utility of the data for AI & Machine Learning models.

So the question is: what’s the best way to implement differential privacy in such a way it won’t minimise the utility of AI & Machine Learning models.

About Info Support Research Center

We anticipate on upcoming and future challenges and ensures our engineers develop cutting-edge solutions based on the latest scientific insights. Our research community proactively tackles emerging technologies. We do this in cooperation with renowned scientists, making sure that research teams are positioned and embedded throughout our organisation and our community, so that their insights are directly applied to our business. We truly believe in sharing knowledge, so we want to do this without any restrictions.

Meld je aan voor deze opdracht

  • Geaccepteerde bestandstypen: docx, doc, txt, pdf.

Sollicitatieprocedure

1

Kennismakingsgesprek

Bespreken (studie)loopbaan, interesses en ambities en introductie Info Support.

2

Assessments

Toetsing vakinhoudelijke kennis en persoonlijkheid (capaciteit, competenties en drijfveren).

3

Selectiegesprek

Verdiepen van vakinhoudelijke kennis en persoonlijkheid.

4

Contractondertekening

Contractaanbieding en uitnodiging tekenmoment.

Andere afstudeeropdrachten

afstudeeropdracht

Text style transfer using deep neural networks

We’ve all seen how neural networks can be used to transfer the artistic style of one image to another image.

afstudeeropdracht

Differential privacy for AI & ML models

Although the implementation of privacy regulations and laws is not a recent development, new laws are being implemented to the challenges and characteristics of the modern information era.

afstudeeropdracht

Automatic assessment of dataset quality for ML

With the rise of automated machine learning tools it becomes much easier for developers to build machine learning models.

afstudeeropdracht

Improve ML text classification models with ensemble learning

Info Support is doing a lot of work for customers that use chatbots to automate customer service processes.