- Experiment: add layers
- Implement and Experiment: Change non-linearities
- Implement: Remove stop-words, weight word vectors by TF-IDF
- Experiment: Dimension of word vectors
- *** Implement and Experiment: Headline/Body Cosine Similarity Feature (append to preds in add_pred_op) ***
- Experiment: do not train word vectors
- Experiment: train word vectors after N epochs, N a hyperparameter
- Experiment: tune batch size
- Implement and Experiment: tune learning rate
- Implement and Experiment: change optimizer
- Implement and Experiment: add regularization
- Experiment: add dropout
- Print training error (after N epochs)
- Verify that scoring in model.py is correct