Applying NLP appropriately

In this unit$$s research task, you investigated natural language processing $($NLP$)$ use cases and related them to the considerations raised by Professor Regina Barzilay. There are two ways to apply natural language processing tools:$$

Option $1$: Train a standard NLP model using large quantities of correctly annotated data.

Option $2$: Use domain experts and NLP experts to build a customized model that can make predictions based on limited data.

Based on your chosen examples for the research task, or an example from the health care sector, which of these two options was most likely used? Support your answers by addressing the considerations raised by Professor Regina Barzilay:$$

Annotation: Is the available data structured and annotated, or will manual annotation be required?

Amount of data: Is there a large amount of relevant data available, or will more data only be accumulated over time?

Retraining: Will the model require subsequent retraining, or is a trained model transferrable across applications?