Which is not a field of action to apply Natural Language processing ?

• Speech Recognition

• Topic extraction from text

• Predict ticket volume at a service channel

• Generate articles or summary documents

Which are typical problems when working with text as data?

• Unstructured, no easy-to-use data model

• some people write uppercase, some lower case

• only feasible on super computers with newest GPUs

• text has to be internally represented as numbers

Mark all typical use cases for NLP.

• Machine Translation

• Sentiment Analysis

• Information Extraction

• Chatbot

Which are reasons to know or detect the language you are working in, although the machine transforms everything into numbers?

• You don't need to, there is one model that fits all languages.

• Every language follows individual grammar and morphology rules

• Only the verbs are relevant, because they differ between the languages.

• There is ambiguous words in different languages that are used in different context.

FastText, BERT, GPT-3 are all examples of...

• Toyota car models

• fancy names of robots

• Neural Networks for language processing

• programming languages for NLP tasks

Amongst all the strength in deep learning models, what is the weakness?

• They can only work with a limited amount of text

• They cannot generalize for different text types

• They are mostly black-boxes and hard to analyse

• Their precision decreases, the more examples you validate