The Responsible AI mentor programme, organised by New York University’s Center for Responsible AI, was launched in 2022. The programme provides Ukrainian university students with the opportunity to participate in a five to ten-month research project in this field, under the guidance of mentor researchers. The main organizer is Julia Stoyanovich. Zsolt Zombori, a scientific researcher at the HUN-REN Alfréd Rényi Institute of Mathematics (HUN-REN Rényi Institute), along with his Oxford colleagues Michael Benedikt and Sam Adam-Day, joined the programme in September 2023.
In early September, during the preparatory phase, the applying mentor groups suggested research topics, which they presented to the Ukrainian students. Subsequently, smaller groups of students interested in the proposed topics – in the case of Zsolt Zombori and his colleagues, approximately 20 individuals – introduced themselves to the mentor in 5 minutes each.
Zombori Zsolt and his colleagues have taken on the mentoring of four computer science students, all of whom are attending the Ukrainian Catholic University, one in Kyiv and the other three in Lviv. They maintain weekly contact with their mentors online. The project's objective is theoretical research, specifically examining how to incorporate constraints in the form of logical formulas into learning systems. Since the behaviour of learning systems is shaped by the available data, providing any guarantee about the final system is very challenging, raising security and ethical issues in many situations. The project aims to discover learning methods that can effectively consider external logical constraints, focusing on two specific datasets, MnistAdd and Road-R. The project requires processing theoretical articles, programming, and conducting simulation experiments.
In case of outstanding collaboration, the joint work can continue.