DL4Seq

Code Assignments of Deep Learning for Text and Sequences (89-687): http://u.cs.biu.ac.il/~89-687/

Assignment 1: Gradients and SGD

• The code was written in python 2.7, using the numpy package.
• Implementation of a log-linear classifier and a multi-layer perceptron classifier.
• The classifiers were tested on the language identification task, and the XOR task.

Assignment 2: Window-based Tagging

• The code was written in python 2.7, using the DyNet package.
• An Implementation of a sequence tagger with window-based features and embeddings.

Assignment 3: RNN Acceptors and Transducers

• The code was written in python 2.7, using the DyNet package.
• An implementation of a squence tagger with LSTM.
• Generate different examples to make the LSTM succeed/fail.