June 10: Dr. Schuman Lecture

 Machine Learning

Give a summary of the information presented and how you think you might use that information in your own schooling.  

Abstract: Today, instead of having a regular split class of STEM Skills and STEM in Society, we had an interactive lecture from Dr. Schuman who is an Assistant Professor in the Department of Electrical Engineering and Computer Science here at UTK. 

During the lecture she began by asking us the first question she asks her students at the beginning of the year. That is "what is machine learning?". To answer this, we essentially came to the conclusion that machine learning is when AI or technology produces tests and data for itself to learn from. Just as we students are trialed and quizzed in order to measure how well we are learning. Scientists and researchers are doing the same for machines and AI. To show us this and how it works even in our daily life Dr. Schuman presented 3 interactive activities for us. 

The first activity involved a website similar to the game many of us have played before: Quick Draw.

 

The only slight difference is that instead of you drawing the item completely you can draw a part of it and AI will finish the drawing based on previously gathered data of how the item looks like and how it was drawn before. I would defiantly say this was an interestingly funny activity as we compared how well the AI did with various items and while some were very good and accurate each time, others... not so much. 
 

The second activity included a site that uses AI to differentiate between two items in which we can see whether it accurately distinguishes between the two and compares the similarity of them. This test was a coin toss. While it had worked well for me, my friend's trial had shown some racial bias discussed a few days before where it had an easier time identifying me, someone with paler skin, compared to her who has a darker skin color. So, it was interesting to see how the AI varied in its performance depending on who the subjects were, and I wonder if there would have been such a significant bias if a man was also a trial subject to compare as well. 

Lastly, Dr. Schuman introduced us to her research now with self-driving cars in which she and her previous class and research group had and still are creating a car that is learning to navigate and bypass winding roads and obstacles on its own without a human controlling it. It is a process in the making but as more and more tests are done on it. It gains more data to apply and teach itself in order to drive itself better.