Cuestionario: AI Language Interaction and Technologies — 9 preguntas

1. What is the primary goal of natural language processing (NLP) in human–computer interaction?

A. To replace human communication entirely

B. To develop new programming languages for AI systems

C. To enable machines to interpret, understand, and generate human language naturally

D. To improve hardware performance for AI applications

Answer: (C)

Explicación

The main goal of NLP within human–computer interaction is to enable machines to interpret, understand, and generate human language in a natural way, making interactions more intuitive and effective.

2. Which of the following models is known for its encoder-only architecture that is capable of understanding and generating language, and has been mentioned in the revision sheet?

A. BERT

B. GPT

C. T5

D. Transformer XL

Answer: (A)

Explicación

BERT is an encoder-only model designed for understanding language, unlike GPT which is decoder-only. T5 is encoder-decoder, and Transformer XL is a transformer model with extended context capabilities. BERT's architecture makes it particularly suitable for tasks requiring deep understanding.

3. Which step is NOT part of the classical NLP pipeline?

A. Image recognition

B. Tokenization

C. Syntax parsing

D. Morphology & POS tagging

Answer: (A)

Explicación

Image recognition is not part of the classical NLP pipeline, which focuses on analyzing and understanding text through steps like tokenization, morphology, syntax parsing, semantics, and pragmatics.

4. What is the primary role of tokenization in the NLP pipeline as described in the revision sheet?

A. To convert text into structured logical formulas

B. To split text into units like words or subwords

C. To analyze grammatical correctness of sentences

D. To generate summaries of the text

Answer: (B)

Explicación

Tokenization is the process of splitting raw text into smaller units such as words or subwords, which is the first step in most NLP pipelines. It does not involve logical formulas, grammatical analysis, or summarization.

5. What is a key advantage of contextual embeddings like BERT and GPT over static embeddings such as Word2Vec?

A. They require less training data

B. They are less complex to implement

C. They dynamically adjust word representations based on surrounding context

D. They are faster to compute

Answer: (C)

Explicación

Contextual embeddings like BERT and GPT dynamically adjust the representation of words based on their surrounding context, allowing better handling of polysemy and context shifts, unlike static embeddings which assign a single vector to each word regardless of context.

6. Which tool mentioned in the revision sheet is known for its speed and rule-based processing?

A. Hugging Face

B. spaCy

C. TensorFlow

D. NLTK

Answer: (B)

Explicación

spaCy is highlighted in the revision sheet as a fast, rule-based NLP tool. Hugging Face is known for state-of-the-art neural models, while TensorFlow and NLTK are different frameworks or libraries for machine learning and NLP, respectively.

7. Which of the following statements accurately describes static word embeddings?

A. They generate different vectors based on context for each word occurrence

B. They, like Word2Vec and GloVe, assign a fixed vector to each word regardless of context

C. They are dynamically generated based on sentence structure and syntax

D. They are only used in rule-based NLP models

Answer: (B)

Explicación

Static embeddings such as Word2Vec and GloVe assign a fixed vector to each word, regardless of the context in which the word appears. This contrasts with contextual embeddings like BERT, which vary based on context.

8. According to the revision sheet, which challenge in NLP involves understanding language that varies depending on the situation and speaker intent?

A. Ambiguity

B. Context-dependence and pragmatics

C. Syntax structure complexity

D. Energy consumption of models

Answer: (B)

Explicación

Context-dependence and pragmatics involve understanding language that relies on context and speaker intent, making interpretation more complex. Ambiguity involves multiple possible meanings but not necessarily the influence of context.

9. Which future trend in NLP, as suggested in the revision sheet, involves improving models by combining retrieved information to augment their knowledge base?

A. Instruction tuning

B. Retrieval augmentation

C. Multilingual models

D. Safety-focused models

Answer: (B)

Explicación

Retrieval augmentation involves enhancing models by retrieving relevant information from external sources, thereby expanding their knowledge and improving performance on various tasks.