9th PhD Workshop already has a winner: Alexandros Zacharakis

9th PhD Workshop already has a winner: Alexandros Zacharakis

On June 4, the 9th edition of the ETIC PhD Workshop was held, with a new format due to the pandemic situation. 19 videos were presented, of which 6 finalists came out.

18.06.2021

After the vote of those attending the event, this year's winner was Alexandros Zacharakis, doctoral student in the Wireless and Secure Communications research group (WISECOM), under the supervision of Dr. Vanesa Daza and Dr. Carla Ràfols.

The winner has been awarded a Dissemination Grant of up to 1400€ to attend a scientific event related to his PhD research.

Alexandros based his doctoral thesis on the problem of computational delegation in the field of computer science, here a more detailed explanation:

Delegation of computations: assumptions vs efficiency

A weak client (for example a smartphone or an IoT device)
wants to delegate the computation of a program P on input x to an
untrusted, computationally powerful server. The server should return the
result of the computation y=P(x) to the client. Since the server is
untrusted, it is also required to send a small proof that allows the
client to verify the claim y=P(x). This should be efficient in the
following sense:

(1) The proof should not be much costlier to compute that executing the
program P.

(2) The proof should be as small as possible.

(3) The verification of the proof should be done much faster than
running the program P.

The first requirement states that the server should not do much more
work than computing the program, while the other two that the client
-which is very weak- can do the verification itself although it cannot
do the computation.

To achieve delegation in practical scenarios we need to make some
assumptions, namely assume that some mathematical problems are hard to
solve, for example the factorization of large integer numbers. One can
rely on many different assumptions, but some are weaker than others.

Ideally one would want to achieve the best efficiency under the weakest
possible assumptions. However, it is not clear what the trade-off is. In
this PhD we investigate this trade-off. We demonstrate that one can
build delegation with a family of very weak assumptions but not with
optimal efficiency. In future work we plan to investigate (1) if it is
possible to use relatively weak assumptions to get (almost) optimal
efficiency and (2) if there is some inherent relation between
assumptions/efficiency.

You can see here the expository video of his thesis.

From the ETIC area we congratulate Alexandros for his work.