Participant Blog: Mykhailo Ivashchenko Masters AI

Mykhailo speaking at the 2019 Sikorsky Challenge. Photo credit: Sikorsky Challenge.

Mykhailo Ivashchenko is an Edmund S. Muskie Professional Fellowship Program alumnus and Fulbright Scholar from Ukraine currently completing a master’s degree at the University of Nebraska-Lincoln (UNL). This summer he interned with the Nebraska Intelligent Mobile Unmanned Systems (NIMBUS) Laboratory at UNL. We caught up with Mykhailo to talk about his interest in artificial intelligence (AI), the research on neural networks verification he completed over the summer, and his reflections on his internship experience.

Tell us about your professional journey up to this point.

I am a computer science engineer focusing on AI, machine learning, and computer vision. I initially chose this field because I was fascinated by everything autonomous. Later on, this interest transformed into a realization that working in this field could improve the well-being of my home country.

During my time in Ukraine, I was part of two computer vision-oriented start-ups. Both were selected as winners at the Sikorsky Challenge, a festival highlighting innovative projects each year in Kyiv. The first start-up, ShotAnalyzer, designed a system for remote, automatic identification of target hits during shooting training. The other one, Deaf Translator, was developing a sign language translation software, but the development was unfortunately shut down due to the COVID pandemic.

What contributed to your interest in interning with your host organization?

Over the summer I interned with the NIMBUS Laboratory. The lab is funded by various institutions around the U.S., such as the Air Force Office of Scientific Research and the National Science Foundation. Its main focus is on autonomous projects like autonomous driving, adaptive sensing, mission planning, and underwater robots.

This work interests me because I’ve always been fascinated with “smart” systems, even ones that involved basic-level programming. As time passed, I started paying attention to systems like smart houses, automated drones, et cetera. My dream is to make an impact by designing robots that would be able to think completely by themselves, also known as “strong AI.”

I decided to intern with the NIMBUS Laboratory as I am interested in working with neural networks in autonomous driving projects. Additionally, the team I joined has done an amazing job in the last couple of years, including presenting key papers at top conferences in the field. Not only did I obtain invaluable technical experience during my internship, but I also met a diverse team of professionals from all around the globe.

The Schorr Center – a large computing data center used by the NIMBUS Laboratory teams at University of Nebraska-Lincoln. Photo credit: University of Nebraska-Lincoln.

Share some of the highlights from your internship experience. Are there any exciting projects you have been working on or any interesting things you have learned?

During my internship, I worked in a research area called “neural networks verification,” and I focused in particular on the algorithms used in autonomous driving.  A neural network is usually designed to replicate the mathematical processes in the human brain. Generally speaking, a network receives an input, analyzes it, and gives a certain output in response.

In this case, the input is an image captured by a camera on an autonomous vehicle. The output is a set of objects identified by the neural network (signs, traffic lights, pedestrians, for example). The main idea behind this is to ensure that the network can be used for the relevant real-world task—in this case, identifying lanes, crosswalks, and traffic lights to help an autonomous vehicle safely navigate its environment. These types of networks allow for faster information processing.

My task was to use a certain verification method to determine the accuracy of these neural networks’ object identification. I was working with a specific type of artificial neural network designed to fit smaller hardware. This type of hardware is usually represented by small chips that can be inserted into autonomous systems like drones, toy helicopters, and plant-watering systems.

In addition to being exciting by itself, this project also allowed me to write a paper describing the methods we designed that could potentially be presented at one of the top conferences in the field.

One of the biggest highlights of the experience was the way the work process was set up. This included how the general working environment was arranged, the high level of preciseness and productivity everything was carried out with, and the overall team atmosphere. There was a separate workplace that looked like a large desk with shelves and space to work. We could have up to four people working in one room at a time. The rooms were filled with different chips, autonomous toy drones, and wires, which added to the overall environment.

The lab wasn’t always fully booked with people, but I could always find someone working every day when I came in in the morning (although sometimes I was the first one in the building). There was always someone who could help, give a piece of advice, or chat, which allowed me to enjoy the work even more. The whole internship was wrapped up with a dinner attended by all the lab members, concluding the summer and giving a start to a new academic year.

Do you have any tips for getting the most out of the experience?

The first tip I can give is to enjoy the time you have as an intern. When I started my internship, I was lucky to get adjusted quickly and easily because of the environment I was part of. You should realize that if you are chosen for an internship, you were the one out of tens or hundreds of people, so even though the actual process may be hard, one still should embrace it. This was my way of enjoying it – every day, you keep working towards the potential impact you could make in terms of reaching the final goal.

The second tip is to work as hard as you can because your time at the internship is limited. Nobody knows what is going to happen afterwards, but your work will always be rewarded in one way or another. I didn’t expect this to happen, but I received a position in the laboratory and will be doing my master’s and doctoral research with the professor I worked with as an intern!

The third and last tip is to make and preserve connections. In the modern world, connections are one of the most important things. Every time I could not find a solution to a problem I was facing, there was someone from another university working on the same topic I could reach out to. I could also reach out to the people in the laboratory I was working with or professors I had studied under in the past. Seemingly random connections may influence our lives in the same way that unexpected events have changed the course of history.

Mykhailo (top row, fifth from left) with the rest of the 2021 Muskie cohort in August. Photo credit: Cultural Vistas.

How will you apply what you have gained from the Muskie Program and your time in the U.S. going forward?

Firstly, I should say that I’ve met a lot of truly remarkable people both in my Muskie Program cohort and at my internship site. In my opinion, these connections might become some of the most important in my life, and I will be reflecting on the experience for a long time.

I also had an opportunity to gain a broad intercultural experience, having met people from five continents in one summer. This definitely allowed me to reevaluate my perception of how people live in other countries, their cultures, and compare them to my own.

This internship allowed me to grasp what means to be a researcher and make a real impact. It also gave me an impression about how research results can be put into practice. I am confident that being a Muskie Fellow will allow me to achieve my long-term goal of improving my home country’s well-being. Ideally, when I am back home, I would like to start an AI-oriented company focused on implementing automated solutions for people with disabilities. Another option would be implementing public well-being solutions with AI, such as road monitoring systems.

I hope that my experience as a Muskie Fellow and Fulbright Scholar will help to spread the word about how these programs can empower talented and determined professionals in different fields.