In this course we will learn to implement, train, debug, visualize and invent neural network models. We will have both HW and a final project that will involve training a complex recurrent neural network and applying it to a large-scale NLP problem. On the model side we will cover word vector representations, window-based neural networks, recurrent neural networks, long-short-term-memory models, recursive neural networks, convolutional neural networks as well as some very novel models involving a memory component.
The course is based on a similar content course given in Stanford last year.
Specifically:
The course is based on a similar content course given in Stanford last year.
Specifically:
Word Vector representations: word2vec, GloVe, language models, softmax, single layer networks
Neural Networks and backpropagation -- for named entity recognition
Practical Tips: gradient checks, overfitting, regularization and activation functions
Recurrent neural networks for language modeling
GRUs and LSTMs -- for machine translation
Recursive neural networks -- for parsing and other tasks
Convolutional neural networks -- for sentence classification
Memory Networks -- for machine translation, question answering and other tasks