Some projects finished in Udacity Deep Learning Nanodegree Foundation
-
dcgan-svhn: use deep convolutional generative adverserial network to generate house number
-
dog-project: use convolutional neural network to detect dog breed
-
embeddings: generate word embeddings used in the NLP task
-
face_generation: use deep convolutional generative adverserial network to generate human faces
-
first-neural-network: simple feed-forward and back-propagation exercise
-
gan_mnist: use generative adverserial network to generate hand written digit
-
intro-to-rnns: basic recurrent neural network exercise, generate sentence trained from Anna Karenina.
-
language-translation: use many-to-many recurrent neural network to do machine translation.
-
reinforcement: reinforcement learning algorithms, including: dynamic programming, Monte Carlo method, temporal difference method, deep q-learning
-
semi-supervised: use generative adverserial network as a tool to do semi-supervised learning, enforce GAN to learn a label for real class
-
sentiment-network: sentiment analysis using basic neural network
-
sentiment-rnn: sentiment analysis using recurrent neural network
-
seq2seq: recurrent neural network sequence to sequence model
-
tv-script-generation: use recurrent neural network to generate scripts for Simpsons.