knowledge-mapper

Project for developing knowledge maps via deep text embeddings.

Getting started

To start a development server

pip install -r requirements.txt
gunicorn --bind 0.0.0.0:8000 wsgi:app

Contents

• static: contains core files (data, config and html) for app
• notebooks: contain preprocessor for extracting text embeddings using BERT

Note: Use any port as available.

TODO

• Create pre-trained map configs for UMAP and t-SNE (update app to new version)
• Add text bar to automagically match top-k units (in vector space) as per query

Acknowledgements and References

• Official TensorFlow Embedding Projector