Thesis supervisor: Bálint Molnár
Location of studies (in Hungarian): ELTE Faculty of Informatics Abbreviation of location of studies: ELTE
Description of the research topic:
The essential feature of gossip-based algorithms - like the real social rumor - does not require full accuracy, high reliability, and synchronous information transmission, instead it allows for asynchronous information exchange. During the calculation, the partial results are secret, even the participating nodes do not know more than their own data. Everyone can only falsify just by manipulating their own data. Data is stored on different nodes. Each node performs some averaging, arithmetic operations, or adds some noise to the calculations, probability variables, distributions that are filtered out by the arithmetic operations or suitably defined "Diophantine arithmetic" calculations. With gossip-based averaging, the average (the result of some arithmetic calculation, a Diophantine operation) could be publicly visible, symmetrically at every location involved in processing; for example, processing and analyzing Big Data with sensitive data content, in the case of health, bio-medical, demographic etc. data warehouses, databases
The research question is how to achieve secure, reliable, trustable and protected information exchange by exploiting probability theory with the combination of Diophantine arithmetic (see Burgin’s work ) through open networks. The algorithms should prevent identifying any single, participant (institution, person etc.) within large-volume data processing processes, in large distributed networks, data warehouses, etc..
Required language skills: English Recommended language skills (in Hungarian): German Number of students who can be accepted: 1