The main research objectives for the project are the following:
- Analyze security of Machine Learning (ML) models against data modifications, using principles of differential cryptanalysis.
- Analyze security of ML models against attacks that aim to recover parameters of the model (e.g., weights of a neural network).
- Investigate whether it is possible to improve security of ML models by making them more resilient against aforesaid attacks, by using techniques from differential and linear cryptanalysis.
The researcher will work with a team at the Selmer Center in Secure Communication including, among others, Adj. Prof. Vincent Rijmen, Adj. Prof. Stjepan Picek, and Prof. Lilya Budaghyan.
Qualifications and personal qualities
- Applicants must hold a Norwegian PhD or an equivalent degree in informatics or mathematics.
- Applicants must have a genuine interest in AI, and the application letter must be related to cryptography and its applications to security of AI.
- Documented research experience in mathematical tools of cryptography is a requirement.
- Experience in machine learning and security issues of AI is an advantage.
- Applicants must be able to work independently in a structured manner and be able to cooperate with others.
- Applicants must have excellent skills in oral and written English.
Personal and relational qualities will be emphasized. Research experience, ambitions and potential will also count when evaluating the candidates.
Special requirements for the position:
The University of Bergen is subjected to the regulation for export control system. The regulation will be applied in the processing of the applications.
We can offer
- A good and professionally stimulating working environment.
- Position as researcher (code 1109 in the basic collective agreement) and a gross annual salary of NOK 624 500-780 000 upon appointment. Further increases in salary are made according to length of service in the position. A higher salary may be considered for a particularly well-qualified applicant.
- Enrolment in the Norwegian Public Service Pension Fund
- Good welfare benefits
Your application must include
- A brief account of the applicant’s research interests and motivation for applying for the position.
- The names and contact information for two referees.
- CV
- Transcripts and diplomas
- Relevant certificates/references
- List of any works of a scientific nature (publication list)
- Publications (as applicable)
The application and appendices with certified translations into English or a Scandinavian language must be uploaded at JobbNorge
General information
For further details about the position, please contact:
Prof. Lilya Budaghyan at the Department of Informatics, Lilya.Budaghyan@uib.no or +47 55584287
Head of Department, Inge Jonassen inge.jonassen@uib.no, +47 55584713
For HR related questions, contact
HR-adviser, Karoline Christiansen at +47 55 58 47 70 or karoline.christiansen@uib.no.
Diversity is a strength that enables us to solve our tasks even better. UiB therefore needs qualified employees regardless of gender, ethnicity, religion, worldview, disability, sexual orientation, gender identity, gender expression, and age.
We encourage women to apply. If multiple applicants have approximately equivalent qualifications, the rules pertaining to moderate gender quotas shall apply.
The University of Bergen applies the principle of public access to information when recruiting staff for academic positions.
Information about applicants may be made public even if the applicant has asked not to be named on the list of persons who have applied. The applicant must be notified if the request to be omitted is not met.
We encourage applicants with disabilities, immigrant backgrounds, or gaps in their CV to apply. By indicating such circumstances in your application, you may receive favourable consideration. We ensure that at least one qualified applicant from each of these groups is invited for an interview as part of our commitment to inclusivity and equal opportunity.
Further information about our employment process can be found here.
