How do we empower non-technical people to train NLP models and deploy them for solving tasks such as sentiment analysis, named entity recognition, and document classification? By abstracting away the implementation details and focusing on domain knowledge transfer, from experts to models, through simple annotations.
This talk demonstrates how John Snow Labs enables the complete workflow from defining a new model, reusing existing models to pre-annotate documents for a faster pace, active learning during annotation to continuously improve results, model evaluation, and finally model publishing. All you need to bring is the experts to learn from and their knowledge materialized as high-quality training data.