InformationExtraction

This is the repository that contains the implementation of a deep learning based natural language models, using 4 types of word embeddings, namely GloVe, ELMo, char-CNN and BERT. Note that the use of the repository is soley on the purpose of the author final year project, Exploratory action selection for discovering object properties.

This repository is structured as followed:

• ./data contains all the data as .csv files, which are used at the time that the models were implemented
• glove_elmo.ipynb contains the implementation of GloVe/ELMo embedding
• char_cnn.ipynb contains the implementation of char-CNN embedding
• BERT.ipynb contains the implementation of BERT embedding

Note that all of the experiment is run using colab GPU