Joshua Steiner's Practicum Observation and Analysis

I have been working as an undergraduate Research Assistant for the DSPCAD research group, supervised by engineering professor Shuvra Bhattacharyya. With this group we have been conducting research on human detection with UAVs and creating a machine-learning algorithm to improve the detection of humans specifically in disaster and war zones, while also being able to detect poses in addition to the presence of a human.

A typical week of work for this project always had me keeping up to date with world news. I was in charge of collecting data from real-life drone footage of disaster areas and war zones, so in order to get the best quality images and build up a dataset of images to work off of, I had to know what was going on in the world. So a lot of research was involved, but then I would actually go on to the internet and find videos on these topics, and upon finding certain frames in the videos that would be good to add to the data set, I would annotate the image. Ultimately I have built up a data set for machine learning as there hasn't been one created for this particular topic before. I utilized OpenCV with Python, Roboflow, and YouTube as my main tools for conducting my work.

By being a part of this project I have learned a lot about AI, machine learning, collecting data, and performing research to do something that hasn't been done before. A concept from the SGC colloquium that reinforced prior knowledge I had was the scientific process. In order to figure out how to go about my work and what would work and what wouldn't work, the scientific process was really important and something that I am still utilizing to this day. For example, I have to identify a problem which is what topics I want to research for footage, and then try searching different phrases online to find footage and go through the data (YouTube videos) to figure out which videos work. And sometimes I don't find a good video. We are looking for specific types of data and many videos do not work or even are titled 'Drone Footage' but do not have any footage in them. There is a lot of trial and error until I eventually find what I am looking for. And then when it comes time for annotations I have to make sure the frames will actually be beneficial for the data set or if I should scrap them and keep looking for different ones. Skills that I have gained from this experience are programming with Python, researching modern topics, patience, and a better conceptual understanding of machine learning.

As I have been doing this I realized this topic of machine learning and different work that the PhD students are doing really interests me and since it is related to Signal and Communications engineering, that is what I am going to specialize in for my Electrical Engineering degree. If/when I pursue graduate school for a Masters degree, Signals and Communications will be the pathway that I take. I wanted to get a better idea of what I wanted to do in the future and this research job has definitely helped me do that since it helped show me what I am interested in. I could have better prepared for this opportunity with having taking more classes relating to machine learning because I have had to learn a lot as I conduct my work. But at the same time because I wasn't as prepared as I could have been, I have had to work really hard and that has had a positive benefit on my character. My next steps for this project are to continue it throughout the summer (and the rest of college as long as our work is still funded), and to increase my involvement in the research. I think this research position will have a positive impact on my professional development as I have had to give presentations and communicate with people higher up than myself in terms of position and knowledge, which is what I will have to do in the real world at a corporation.

As a part of my practicum, I had to attend weekly Zoom meetings and discuss with everyone what my progress has been so far and see what everyone else is talking about. Most of my work involves me performing my own tasks, but I keep everyone informed on what I am doing so that they are aware. I also send weekly reports to my supervising professor. My work is more or less independent, but it is very related to what everyone else is doing so I get a lot of information from our weekly meetings. I have never been in this kind of role before and it has been a very interesting experience.

I would recommend this opportunity to future SGC students as it is related to a very interesting topic and has enabled me to know what kind of work and field I want to go into in the future. It has also sparked my own interest in these topics and has helped me develop professionally, especially when giving a presentation to representatives from the Army Research Lab. Really, choose a practicum project that you think you will find interesting and it will probably go well and you will learn a lot from it.