Question answering incorporating knowledge graphs Milestones: 28/02/2022
- Produced knowledge graph from input data
- Started testing over different tools to learn the embeddings of the graph (PyKeen, node2vec)
- Produced word embeddings from input data using a SkipGram model (A better corpus should be used when hardware is available)
- Produced QA Tensorflow model using a EncoderDecoder + attention Pattern (Should later replace attention with the anchor loss to add context)
- AnchorLoss written and implemented
- Knowledge graph trained and emmbedded
- Whole pipeline and it's testing implemented
--TODO
- Properly train the model
- Validate the model