QCM : AI Language Interaction and Technologies — 9 questions

Questions et réponses du QCM

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

To replace human communication entirely
To develop new programming languages for AI systems
To enable machines to interpret, understand, and generate human language naturally
To improve hardware performance for AI applications

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

Explication

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?

BERT
GPT
T5
Transformer XL

BERT

Explication

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?

Image recognition
Tokenization
Syntax parsing
Morphology & POS tagging

Image recognition

Explication

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?

To convert text into structured logical formulas
To split text into units like words or subwords
To analyze grammatical correctness of sentences
To generate summaries of the text

To split text into units like words or subwords

Explication

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?

They require less training data
They are less complex to implement
They dynamically adjust word representations based on surrounding context
They are faster to compute

They dynamically adjust word representations based on surrounding context

Explication

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?

Hugging Face
spaCy
TensorFlow
NLTK

spaCy

Explication

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?

They generate different vectors based on context for each word occurrence
They, like Word2Vec and GloVe, assign a fixed vector to each word regardless of context
They are dynamically generated based on sentence structure and syntax
They are only used in rule-based NLP models

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

Explication

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?

Ambiguity
Context-dependence and pragmatics
Syntax structure complexity
Energy consumption of models

Context-dependence and pragmatics

Explication

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?

Instruction tuning
Retrieval augmentation
Multilingual models
Safety-focused models

Retrieval augmentation

Explication

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

Révisez avec les flashcards

Mémorisez les réponses avec 10 flashcards sur AI Language Interaction and Technologies.

NLP — core task?

Extract meaning from human language

NLP — definition?

Enables machines to interpret and generate human language.

Classical NLP pipeline — step?

Tokenization, morphology, syntax, semantics, pragmatics

Voir les flashcards →

Approfondir avec la fiche

Consultez la fiche de révision complète sur AI Language Interaction and Technologies.

Voir la fiche →

Cours similaires

Introduction à l'Informatique et IA

Introduction aux marchés publics et réglementations du bâtiment

Bases de données

Gestion des imprimantes Windows et PDF

Introduction aux bases de données relationnelles

Bases de données

Maîtrise des Boucles en Programmation

Programmation

Maîtrise des fonctions et paradigmes de programmation

Programmation

Crée tes propres QCM

Importe ton cours et l'IA génère des QCM avec corrections en 30 secondes.

Générateur de QCM