Distributed Multipoint Functions from Distributed Interval Functions
Organisatsiooni nimi
Cryptography Research Group
Kokkuvõte
Secure Multiparty Computation is a technique that allows a set of parties to evaluate a function while not learning anything about the inputs. A common approach is the so-called offline-online approach, where a number of correlated randomness is generated beforehand that allows to make actual computation much faster. In the recent years there has been a lot of research focused on correlated (pseudo)randomness generators, which are techniques which would allow to speed up the correlated randomness generation a lot.
\t\t
These are all based on a technique called Distributed Point Functions, where two parties can jointly, without communication, evaluate a function that is nonzero at a specific point, and zero everywhere else. This type of function is then evaluated, say, a thousand times, for different points to build the Pseudorandom Correlation Generator. In essence, they build a Distributed Multipoint Function from single Distributed Point Functions.
\tOne of the same papers that proposed the original construction for a distributed point generator also proposed something called a Distributed interval function, which evaluates to nonzero at a given, secret, interval, and zero everywhere else.
\t\tThe masters thesis would involve studying an idea to see whether we could use the Distributed Interval Function to build a Distributed Multipoint Function.
\t\t
These are all based on a technique called Distributed Point Functions, where two parties can jointly, without communication, evaluate a function that is nonzero at a specific point, and zero everywhere else. This type of function is then evaluated, say, a thousand times, for different points to build the Pseudorandom Correlation Generator. In essence, they build a Distributed Multipoint Function from single Distributed Point Functions.
\tOne of the same papers that proposed the original construction for a distributed point generator also proposed something called a Distributed interval function, which evaluates to nonzero at a given, secret, interval, and zero everywhere else.
\t\tThe masters thesis would involve studying an idea to see whether we could use the Distributed Interval Function to build a Distributed Multipoint Function.
Lõputöö kaitsmise aasta
2024-2025
Juhendaja
Toomas Krips
Suhtlemiskeel(ed)
eesti keel
Nõuded kandideerijale
This is a MsC topic. It requires some knowledge in Cryptography, such as having taken an introductory course in cryptography.
Tase
Magister
Kandideerimise kontakt
Nimi
Toomas Krips
Tel
E-mail