top of page
Search
  • arodri393

Demystifying Machine Learning

Updated: May 12, 2021



When you think of Machine Learning, what comes to mind? Face ID, artificial intelligence or robots taking over the world? Well fortunately, not all machine learning is as intimidating as it may seem. It can be very useful to analyze data in various career settings like; medicine, business, engineering, the sciences etc. Sometimes, data can be overwhelmingly large and it is difficult to go through it without the help of a computer. For this project, machine learning turned out to be the answer to our prayers! It allowed us to distinguish between two biological organism; trees and flowers.


We collected about 80 total pictures of trees and flowers and we needed to separate them into two categories, which I know doesn't seem to be a lot, but why not allow the computer to do the hard work for you, right? Well that's exactly what we did.


Flowers 🌸












Trees 🌳









With the help of Dr. Ilmi Yoon's brilliant image classification code, we were able to use a program called "VGG16 model" to train our computer to predict whether the image at hand was a tree or a flower. The program runs by training, then validating and finally by testing the images.


The code seemed to effectively train and validate all images. However, just like all programs it isn't perfect, so it incorrectly predicted one image out of twenty. Below are some pictures of our results:



Prediction was 100% accurate.













Prediction was 100% accurate.













Prediction was not correct, as it predicted it was a "flower" when it is a "tree".











So next time, when you're stressed about having an overwhelming amount of visual data to go through think about using machine learning to get your work done!

39 views0 comments
bottom of page