Virtual Labs

Building Chunker

What is a 'chunk' in the context of Natural Language Processing?

a: A non-overlapping, meaningful group of words such as a noun or verb phrase Explanation

Explanation

b: A single word only Explanation

Explanation

c: A random group of words Explanation

Explanation

d: A full syntactic tree Explanation

Explanation

Which of the following is NOT typically used as a feature for chunking?

a: Word itself (lexicon) Explanation

Explanation

b: Part-of-speech (POS) tag Explanation

Explanation

c: Font color Explanation

Explanation

d: Neighboring POS tags Explanation

Explanation

What is the main advantage of using both lexicon and POS features for chunking?

a: It provides richer information, improving chunking accuracy Explanation

Explanation

b: It makes the model slower Explanation

Explanation

c: It reduces the number of chunks Explanation

Explanation

d: It is required for all languages Explanation

Explanation

If you increase the size of the training corpus for a chunker, what is the most likely effect?

a: Accuracy increases (up to a point) Explanation

Explanation

b: Accuracy decreases Explanation

Explanation

c: No effect on accuracy Explanation

Explanation

d: The model stops working Explanation

Explanation

Which model is generally more flexible for chunking: HMM or CRF?

a: HMM Explanation

Explanation

b: CRF Explanation

Explanation

c: Both are equally flexible Explanation

Explanation

d: Neither can be used for chunking Explanation

Explanation

What is a common error type in chunking output?

a: Boundary error (incorrect chunk start/end) Explanation

Explanation

b: Font error Explanation

Explanation

c: POS tag error only Explanation

Explanation

d: No errors occur in chunking Explanation

Explanation

Suppose your chunker is making many mistakes. Which of the following is LEAST likely to help?

a: Adding more labeled training data Explanation

Explanation

b: Using richer features (lexicon + POS) Explanation

Explanation

c: Randomly shuffling the training labels Explanation

Explanation

d: Switching from HMM to CRF Explanation

Explanation

What does the 'Check Accuracy' step in the simulation demonstrate?

a: How well the chunker performs on unseen data Explanation

Explanation

b: How to count the number of words in a sentence Explanation

Explanation

c: How to translate a sentence Explanation

Explanation

d: How to generate random sentences Explanation

Explanation

Which of the following best describes the difference between chunking and full parsing?

a: Chunking finds non-overlapping groups, full parsing finds complete syntactic structure Explanation

Explanation

b: Chunking is slower than full parsing Explanation

Explanation

c: Full parsing ignores word order Explanation

Explanation

d: Chunking is only used for English Explanation

Explanation

In practical NLP applications, why is chunking useful?

a: It helps identify meaningful phrases for downstream tasks like information extraction Explanation

Explanation

b: It replaces the need for all other NLP tasks Explanation

Explanation

c: It is only used for text generation Explanation

Explanation

d: It is not useful in practice Explanation

Explanation