Alexander Wu's repositories
ANN_from_scratch_no_libraries_just_math
To get back in touch with the fundamentals of AI, I built a vanilla neural network using nothing but python. No libraries, just math. Not even numpy.
Artificial-Neural-Network-Bank-Classifier
A classifier that determines whether a user will stay with a bank according to their bank account information
Autoencoder
An autoencoder recommender system. It will predict a user's 1-5 rating of a new movies based off their ratings of previously watched movies.
CNN_Image_Classifier_Pytorch
I built a CNN to classify images of flowers into 102 distinct species. I use transfer learning from densenet161, and added my own trained 3 fully connected classifier layer. This is all done using the Pytorch library.
Convolutional_Dog-Cat_Classifier
A convolutional neural network that predicts whether a new image is either a dog or a cat
Firmware
PX4 Autopilot Software
Recurrent_LSTM_Stock_Prediction
Using the past 2 months of daily opening prices for GOOGL, the RNN would predict the next day's opening price
Restricted_Boltzmann_Machine
A manually built restricted boltzmann machine recommender system. It looks at a customer's movie preferences and decides wether or not they will like a new movie.
Self_Organizing_Map_Feature_Detection_With_ANN
A self organizing map that uses customer data to identify fraudulent customers. Then, the entire data is used to train a vanilla neural network to identify fraudulent behaviour.
TicTacToe-Ception_ReactJS
First project using ReactJS - I made a tictactoe game within a tictactoe game