TextDistil - A Robust
Knowledge Extraction Pipeline
Built on LLMs and Semantic Technology
TextDistil in Language and Knowledge techscape
TextDistil targets the cognitive corner of the technology land scape. It is well positioned to deliver novel functionality by leveraging the amazing power of Large Language Models combined with the robust methods of Semantic Technology.
Accurate Entity Detection
Beside open domain entities, our models are tuned for domain specific taxonomy and trained on large curated domain specific data sets.
Exploit fully the context in narrative Text
Entity and Relation prediction in narrative text is further enhanced with our purpose built pragmatics augmented co-reference models.
Leverage the power of LLMs and Semantic Technology
We leverage the awesome power of Large Languge Models (GPT3.5) to tease out the understanding to construct high fidelity domain specfic knowledge as described by an Ontology using Semantic Technology.
Configurable Cognition Pipeline
Architecture of TextDistil pipeline lends it self to be domain independent. The pipeline enables ‘no code’ Knowledge Graph generation from Text and adapts to different domains and down stream applications by configuring custom Ontology and custom Taxonomy trained models.
