- Create novel word embeddings which capture maximum semantic information and entropy.
- Input this word embeddings to the encoder
- Create a basic encoder-decoder model using these embeddings for test.
- Develop and train a RNN Encoder-decoder model using Glove vectors word embeddings.