 |
| Maaz and his partner developing their project |
The capstone experience was surely a unique one for me. I still remember taking HON 301 with Prof. Edward Miller in my junior year when I had no idea what I was going to explore. The class taught me not only about research but also the whole experience surrounding it; it gave me confidence about working on new research that even my faculty advisor hadn’t started yet. As I finished the class, I started working on my actual research project with Prof. Feng Gu and my research partner, Tatiana Anderson. The project transformed throughout the time span of a year, and communication on each step played a vital role as being on the same page is very important.
The topic of my research is Predictive Policing. My partner and I separated the jobs of data gathering and machine learning between us and met weekly. One of the first milestones that I covered was when I successfully implemented Google API according to our needs and automated the process of using API for more than 10,000 coordinates. A lot of testing and research went into gathering data. The second source we used for data gathering was Census, which is a great tool when used for a specific area (Census Tract in our case). Their API didn’t have as much documentation as the Google API did, and after a lot of dead-ends we were able to implement that for specific variables too.
I was always fascinated by crime-related TV shows and movies, because it would portray how different tactics are considered by both sides. Predictive policing is currently in the infancy stages, but it is already becoming a tool to revolutionize policing. Every new technology has its pros and cons, nevertheless predictive policing is being used by police departments in cities such as NYC and Chicago. It was reported that an algorithm using big data analysis resulted in a 80% accuracy about the individuals most likely to get shot on Mother’s Day weekend of 2016 in Chicago, IL. Although the tragic news that nearly all the victims who were predicted did get shot, the confidence in predictive policing has been on the rise as well. A lot more research is still needed but if we can keep digging for the best results, an algorithm that has the potential to save lives isn’t far away. During the Fall semester of 2017, an opportunity to socialize and make connections was presented by Prof. Feng Gu. In this social event, I met multiple PhD students and CSC alumni who had worked on research with Prof. Gu. The research experience felt at times as we were coworkers of professor because of the respect that was given to us throughout that evening by other students and professionals. I not only felt comfortable researching but also learned a lot about writing, presenting and interacting with professionals throughout my research experience.
References:
https://www.smithsonianmag.com/innovation/artificial-intelligence-is-now-used-predict-crime-is-it-biased-180968337/
http://www.dailyhomerecipes.com/wp-content/uploads/sites/952/2015/05/the-best-grilled-bbq-chicken-kebabs.jpg
No comments:
Post a Comment