Neural Networks and Machine Learning by Griffin Carroll
I am a motivated learner
I am a freshman at the University of Iowa double majoring in Data Science and Physics. I have a deep passion for both fields and have been actively engaged in learning and applying the principles of these subjects. Through my coursework and independent projects, I have gained proficiency in programming, data analysis, and problem-solving. I am excited to continue my education and contribute to the fields of data science and physics.
My Start
As a Junior in high school, I felt like I was starting to understand my interests all until I discovered programming. I began my journey by learning the basics of python. Over the course of many months, I learned datatypes, sequence types, conditions, and iteration. Quickly though, I became enthusiastic about object oriented programming. I realized early on the OOP could have a significant amount of power given the correct context.
About this time, the new ChatGPT 3 model came out and AI technology suddenly became the rage. However, while other people were interested in using AI applications, I was learning to program them. I came to understand that knowing how to program a model would get me nowhere if I didn’t understand the math. Therefore, I spent much of my Senior year in high school as well as the following summer, self teaching Linear Algebra and Calc 3 topics in order to understand what was truly going on. Following that I built my own library and model.
Where I Am Now
In the past few months, I have programmed many models. The types I have programmed include linear regression, image classification, and GAN. Currently I am working on a UNET architecture that is paired with Dice Loss to obtain higher quality segmentation in a more timely manner.
Over the past summer, I worked to understand the math that was going on in neural networks and convolutional neural networks, and because of that, I am able to accurately implement models that fit specifications with few bugs.
Following my understanding of the UNET, I will be adding to the output of the University of Iowa robotics team to automate the process of fulfilling a task based on image recognition and autonomy.