A research team is building a language model for a low-resource language. They have a small corpus of text. Which approach is most likely to improve the model's ability to predict word sequences?

Use transfer learning from a multilingual pre-trained model.

Apply aggressive stemming to reduce vocabulary size.

Use a rule-based system with hand-crafted grammar rules.

Train a statistical n-gram model from scratch on the small corpus.

A financial news analysis firm wants to build a system that can detect market-moving events from news articles. The system must handle sarcasm, irony, and ambiguous language. The team has a small labeled dataset of articles with event annotations. They are considering: (1) using a pre-trained transformer model fine-tuned on the labeled data, (2) using a lexicon-based sentiment analysis tool, (3) using a rule-based system with hand-crafted patterns, (4) using a simple bag-of-words model with TF-IDF. Which approach is most likely to handle the nuances of sarcasm and ambiguity effectively?

Fine-tune a pre-trained transformer model on the labeled dataset.

Use a bag-of-words model with TF-IDF.

Use a rule-based system with hand-crafted patterns.

Use a lexicon-based sentiment analysis tool.

A voice assistant company wants to improve its speech recognition system by predicting the most likely word sequence given an audio signal. Which type of model is best suited for this task?

A stemming algorithm that reduces words to root forms

A stop-word removal process that filters common words

A statistical language model that assigns probabilities to word sequences

A bag-of-words model that counts word frequencies

A hospital's NLP system is used to extract patient symptoms from clinical notes. The system misinterprets 'the patient denies chest pain' as a positive symptom because it only looks at keyword presence. Which NLP challenge is this?

Named entity recognition failure

EC-CouncilArtificial Intelligence Essentials

Domain 3Objective 4

Natural Language Processing (NLP) AIE Practice Questions (Page 1)

Part of the Building Blocks of AI domain, which makes up ~25% of our current practice bank.

38questions here

8free pages

5concepts

Questions 1–5

1application · medium
A research team is building a language model for a low-resource language. They have a small corpus of text. Which approach is most likely to improve the model's ability to predict word sequences?
Select an answer first
2application · easy
A multinational e-commerce company wants to automatically translate customer reviews from Spanish to English. Which NLP application should they implement?
Select an answer first
3expert · hard
A financial news analysis firm wants to build a system that can detect market-moving events from news articles. The system must handle sarcasm, irony, and ambiguous language. The team has a small labeled dataset of articles with event annotations. They are considering: (1) using a pre-trained transformer model fine-tuned on the labeled data, (2) using a lexicon-based sentiment analysis tool, (3) using a rule-based system with hand-crafted patterns, (4) using a simple bag-of-words model with TF-IDF. Which approach is most likely to handle the nuances of sarcasm and ambiguity effectively?
Select an answer first
4application · medium
A voice assistant company wants to improve its speech recognition system by predicting the most likely word sequence given an audio signal. Which type of model is best suited for this task?
Select an answer first
5application · medium
A hospital's NLP system is used to extract patient symptoms from clinical notes. The system misinterprets 'the patient denies chest pain' as a positive symptom because it only looks at keyword presence. Which NLP challenge is this?
Select an answer first

Finished these 5 questions?

Review the revealed explanations, or continue through the curriculum.

Next 5 questions →Back to AIE

Free Basic Practice is a study aid with revealable answers — not a scored exam. Examers.io is independent and not affiliated with or endorsed by EC-Council. “AIE” is a trademark of its owner, used for identification only.

Natural Language Processing (NLP) AIE Practice Questions (Page 1)

Questions 1–5

A research team is building a language model for a low-resource language. They have a small corpus of text. Which approach is most likely to improve the model's ability to predict word sequences?

A multinational e-commerce company wants to automatically translate customer reviews from Spanish to English. Which NLP application should they implement?

A voice assistant company wants to improve its speech recognition system by predicting the most likely word sequence given an audio signal. Which type of model is best suited for this task?

A hospital's NLP system is used to extract patient symptoms from clinical notes. The system misinterprets 'the patient denies chest pain' as a positive symptom because it only looks at keyword presence. Which NLP challenge is this?

Keep studying — Building Blocks of AI