Graduate Student of the Faculty of Computer Science on Studying in a Joint PhD Programme, Complex Systems, and Working During a Pandemic

Photo courtesy of Roman Nesterov

For Roman Nesterov, PhD student and research assistant at the Laboratory of Process-Aware Information Systems (PAIS Lab), working remotely had already been the norm long before this spring. As student of a cotutelle PhD programme, he is constantly working with colleagues from Italy. We spoke with Roman about his research and about how one can pursue joint doctoral studies with another university abroad.

How is your PhD programme structured? How did this programme come to be?

I am in my third year of cotutelle doctoral studies at the Faculty of Computer Science at HSE University and the University of Milan-Bicocca (Università degli Studi di Milano-Bicocca, UNIMIB). My academic advisor at HSE is Professor and PAIS Lab Head Irina Lomazova, and my advisor at the University of Milan-Bicocca is Professor Lucia Pomello of the Department of Informatics, Systems, and Communications (Dipartimento di Informatica, Sistemistica e Communicazione, DISCo).

I am studying in an full time advanced doctoral programme, a mandatory part of which is completing internships at foreign universities. Professor Lomazova suggested taking things further and organizing a full-fledged double PhD programme. The system of double academic advising (co-tutoring) is widely developed in European universities. My advisors agreed upon the form their joint academic supervision would take and what would be expected of me, and this agreement went through all the standard approval procedures. Once the agreement was signed, in my very first year of study, I became a doctoral student of two universities at once. Due to the fact that HSE had switched to a different degree system (PhD HSE) in which PhDs are awarded, arranging this cotutelle programme turned out to be easier than one might expect. Under the joint academic supervision agreement, I need to prepare one thesis text in English and pass a single defense that will be held at HSE.

Photo courtesy of Roman Nesterov

What is the focus of your research?

I am modeling the behavior of complex multi-agent systems, or systems that consist of many components. System agents interact with each other: for example, they send messages to each other or perform synchronous operations. A monolithic approach to modeling such systems does not take into account agents’ individual behaviors, and may consequently be inadequate. Therefore, we are working on a compositional approach, whereby a model of the entire system is composed of models of agents in accordance with how they should interact.

However, this approach may also have disadvantages. For example, a model may be composed of two correct agents (i.e., agents that work without deadlocks and correctly terminate), but ultimately end up with incorrect behavior. It can be extremely difficult to fully validate the model since there may not be enough computational resources. With the help of special local individual agent model checks, we were able to prove that the correctness of the model of the entire system will follow from the correctness of the models of individual agents. This method greatly simplifies the modeling of the behavior of multi-agent systems.

We are applying our approach to solving the problem of synthesizing models from information systems event logs (process mining). Based on real data, we are building well-structured models that reflect the real state of the system: over time, they can change and differ from the form they were originally planned.

Photo courtesy of Roman Nesterov

What form does your work with Milan take?

In accordance with the agreement, I spend most of my time in Russia, and several times a year I travel to Italy and work there, as part of the team of the Models of Concurrency, Computation and Communication Laboratory (MC3). Usually, before arriving, we designate a specific set of tasks that I will deal with, and I devote all my time in Milan to those tasks. Before the pandemic, I traveled to Milan twice a year, each time for a month and a half. My last trip to Milan was actually in January-February 2020. I believe that it is thanks to the integrated approach we took to organizing our joint work that we have been able to obtain interesting results. My trips to Milan are always fruitful and exciting, because, in addition to our lab work, we manage to find time to enjoy Italian cuisine, take endless walks in Milan's parks and, of course, visit unique museum exhibits.

Has working remotely affected your work?

I can say that the quarantine has not had a significant impact on our joint work—we often met online before the pandemic. For example, in May we had to prepare the final version of our article for the international workshop ‘Petri Nets and Software Engineering (PNSE-2020)’, which was held at the end of June. We prepared the article on time, but the conference itself was, of course, online. In addition, in June I underwent a routine performance review at HSE and the University of Milan-Bicocca. It entails a fairly standard procedure: you present your work and plans at a department meeting, and there is a question and answer session.

For me, the only catch with working remotely—when the boundaries between home and work become blurred—turned out to be psychological. But thanks largely to Irina Alexandrovna, Lucia, as well as my friends and colleagues from the Faculty of Computer Science, I managed to overcome all the difficulties of the period of self-isolation, which, by the way, coincided with my birthday. When the opportunity arose to return to ‘offline mode’, I took it — working onsite is more comfortable for me. I tried to organize my work so that the psychological discomfort from being at home for so long did not affect the results of my work, but this does not mean that I managed to eliminate the discomfort entirely.