Using Machine Learning to Recognizing Graphs

Mentor: Christopher McCarthy

Abstract:Machine learning has been applied successfully to many domains, ranging from acoustics, to images recognition and processing, to natural language processing. One focus of this project is to build and train neural networks to distinguish images of mathematical graphs. The coding language used is Python, both on its own, and with the help of machine learning packages such as TensorFlow (by Google) and Keras. So far, our neural net is capable of distinguishing straight lines and parabolas with high accuracy. Further goals are to train the net to distinguish other functions (for example the trigonometric functions).