Spaces:

Seetha

/

IMA-pipeline-streamlit

like 1

Sleeping

This is the demo of our research work https://www.mdpi.com/2078-2489/14/7/367.

This work converts a text document to a knowledge graph. The input to the pipeline will be a pdf file, and the output will be a knowledge graph. The causal statements are extracted and categorized in the pipeline based on a stakeholder taxonomy we defined. A knowledge graph will be created based on the output of the NLP pipeline (extracted causalities that are classified based on stakeholder taxonomy). This app will better work for financial text documents because the model is trained on financial text. It can be extended to any text.
We are using three models for each of the classification and extraction tasks in one pipeline. We need a GPU grant to run this pipeline faster.

Thanks and Regards,
Seetha