Natural Language Processing - Practice MCQs for CCAT

Correct Answer: B - Natural Language Processing

NLP (Natural Language Processing) is AI focused on interaction between computers and human language.

Correct Answer: C - Breaking text into words or subwords

Tokenization splits text into individual units (tokens) like words, subwords, or characters.

Correct Answer: B - Common words like "the", "is" often removed

Stop words are common words (the, is, at, etc.) often removed as they add little meaning.

Correct Answer: C - Root/base form (may not be valid word)

Stemming removes suffixes to get word stems (running → runn), which may not be valid words.

Correct Answer: B - Producing valid dictionary words

Lemmatization uses vocabulary and morphological analysis to return valid base forms (running → run).

Correct Answer: D - Word importance in document relative to corpus

TF-IDF (Term Frequency-Inverse Document Frequency) measures how important a word is to a document in a corpus.

Correct Answer: C - Represents text as word frequency counts ignoring order

Bag of Words represents text as a collection of word counts, ignoring grammar and word order.

Correct Answer: A - Represent words as dense vectors capturing semantic meaning

Word embeddings map words to dense vectors where similar words have similar vector representations.

Correct Answer: B - Named entities like persons, locations, organizations

NER identifies and classifies named entities in text into categories like person, organization, location.

Correct Answer: D - Emotional tone (positive/negative/neutral)

Sentiment analysis identifies the emotional tone or attitude expressed in text.

Correct Answer: B - Labels words with grammatical categories

POS tagging assigns grammatical categories (noun, verb, adjective, etc.) to words in text.

Correct Answer: B - Transformer-based language model

BERT (Bidirectional Encoder Representations from Transformers) is a pre-trained language model for various NLP tasks.

Correct Answer: B - Focuses on relevant parts of input

Attention allows models to focus on relevant parts of input when producing output, weighing importance dynamically.

Correct Answer: C - Text from one language to another

Machine Translation automatically translates text from one natural language to another.

Correct Answer: D - Categorizing documents into predefined classes

Text classification assigns documents to predefined categories like spam detection or topic labeling.

Correct Answer: B - Contiguous sequences of n items from text

N-grams are contiguous sequences of n items (words or characters) from text, capturing local context.

Correct Answer: C - Sequential data like text

RNNs process sequential data by maintaining hidden state that captures information from previous inputs.

Correct Answer: A - Vanishing gradient and short-term memory

LSTM (Long Short-Term Memory) uses gates to preserve long-term dependencies and mitigate vanishing gradients.

Correct Answer: B - Process sequences in parallel with attention

Transformers use self-attention to process all positions in parallel, enabling faster training and better long-range dependencies.

Correct Answer: A - Self-supervised learning on large text corpora

GPT (Generative Pre-trained Transformer) is pre-trained using self-supervised learning to predict next tokens in text.

Correct Answer: C - Breaking text into smaller units like words or sentences

Tokenization is the process of breaking text into smaller units called tokens, which can be words, subwords, or sentences. It is typically the first step in any NLP pipeline.

Correct Answer: D - Root or base form by removing affixes

Stemming reduces words to their root form by removing suffixes and prefixes. For example, 'running', 'runs', 'ran' are all reduced to 'run'. It may produce non-dictionary words.

Correct Answer: B - Identify and classify named entities like persons, organizations, and locations

NER identifies and classifies named entities in text into predefined categories such as person names, organizations, locations, dates, and monetary values.

Correct Answer: A - The emotional tone or opinion expressed in text

Sentiment Analysis determines the emotional tone behind text — whether it expresses positive, negative, or neutral sentiment. It is widely used in analyzing customer reviews and social media.

Correct Answer: A - Lemmatization produces valid dictionary words; stemming may not

Lemmatization uses vocabulary and morphological analysis to return valid dictionary words (lemmas), while stemming uses simple rule-based suffix stripping which may produce non-dictionary words.

Correct Answer: C - Term Frequency-Inverse Document Frequency

TF-IDF (Term Frequency-Inverse Document Frequency) is a statistical measure that evaluates how important a word is to a document in a collection. It increases with frequency in a document but decreases with frequency across documents.

Correct Answer: B - The

Stop words are common words like 'the', 'is', 'at', 'which', 'on' that are filtered out during text preprocessing because they carry little meaningful information for analysis.

Correct Answer: A - A multiset of words disregarding grammar and word order

Bag of Words represents text as a multiset (bag) of words, disregarding grammar and word order. Each document is represented as a vector of word counts or frequencies.

Correct Answer: D - Creating word embeddings

Word2Vec creates dense vector representations (embeddings) of words that capture semantic relationships. Words with similar meanings have similar vectors. It uses CBOW or Skip-gram architectures.

Correct Answer: D - Grammatical categories like noun, verb, adjective to each word

POS tagging assigns grammatical categories (noun, verb, adjective, adverb, etc.) to each word in a sentence. It is fundamental to many NLP tasks like parsing and information extraction.

Correct Answer: A - Speech Recognition

Speech Recognition (also called Automatic Speech Recognition or ASR) converts spoken language into written text. It is used in virtual assistants, dictation software, and voice-controlled systems.

Correct Answer: D - Contiguous sequences of n items from text

N-grams are contiguous sequences of n items (words or characters) from text. Unigrams (n=1), bigrams (n=2), and trigrams (n=3) are commonly used for language modeling and text analysis.

Correct Answer: D - Word Embeddings

Word Embeddings (like Word2Vec, GloVe, FastText) convert words into fixed-length dense vectors that capture semantic meaning. Unlike one-hot encoding, similar words have similar vector representations.

Correct Answer: A - Supervised learning

Text classification is a supervised learning task where the model is trained on labeled text data to assign predefined categories to new text. Examples include spam detection and topic classification.

Correct Answer: C - A large collection of text documents

A corpus (plural: corpora) is a large, structured collection of text documents used for training and evaluating NLP models. Examples include Wikipedia, news articles, and book collections.

Correct Answer: C - Automatically translating text from one natural language to another

Machine Translation automatically translates text from one natural language to another (e.g., English to French). Modern approaches use neural networks (Neural Machine Translation).

Correct Answer: A - Transformer

The Transformer architecture revolutionized NLP with its self-attention mechanism that processes all positions in a sequence simultaneously. BERT, GPT, and T5 are all based on Transformers.

Correct Answer: A - Bidirectional Encoder Representations from Transformers

BERT (Bidirectional Encoder Representations from Transformers) is a pre-trained language model by Google that considers context from both directions (left and right) of a word simultaneously.

Correct Answer: D - Cleaning and preparing raw text data for analysis

Text preprocessing involves cleaning and preparing raw text for NLP tasks. Common steps include tokenization, lowercasing, removing stop words, stemming/lemmatization, and handling special characters.

Correct Answer: A - Group words into meaningful phrases based on POS tags

Chunking (shallow parsing) groups adjacent words into meaningful phrases based on their POS tags. For example, grouping 'the big cat' as a noun phrase (NP). It provides partial syntactic analysis.

Correct Answer: C - Image Segmentation

Image Segmentation is a computer vision task, not an NLP application. Chatbots, spam filtering, and language translation are all common applications of Natural Language Processing.

Correct Answer: C - Grammatical relationships between words in a sentence

Dependency parsing analyzes the grammatical structure of a sentence by identifying relationships (dependencies) between words, such as subject-verb, verb-object, and modifier relationships.

Correct Answer: A - To represent words as numerical vectors capturing semantic meaning

Word embeddings represent words as dense numerical vectors in a continuous vector space where semantically similar words are closer together. This enables mathematical operations on word meanings.

Correct Answer: D - Subword tokenization

Subword tokenization (like BPE, WordPiece) handles OOV words by breaking unknown words into smaller subword units that exist in the vocabulary. This is used in models like BERT and GPT.

Correct Answer: B - Automatically extracting structured information from unstructured text

Information Extraction automatically extracts structured data (entities, relationships, events) from unstructured text. It includes tasks like NER, relation extraction, and event extraction.

Correct Answer: C - Measuring similarity between two text vectors

Cosine similarity measures the cosine of the angle between two vectors, indicating how similar they are in direction regardless of magnitude. It is widely used to compare document or word vectors in NLP.

Correct Answer: D - Machine Translation

Machine Translation is a sequence-to-sequence task that takes an input sequence (source language) and produces an output sequence (target language). Text summarization is another seq-to-seq task.

Correct Answer: A - Sparse, high-dimensional vectors

One-hot encoding represents each word as a sparse, high-dimensional vector where only one element is 1 and the rest are 0. The vector dimension equals the vocabulary size, making it memory-inefficient.

Correct Answer: D - To focus on relevant parts of the input when generating output

The attention mechanism allows the model to focus on the most relevant parts of the input sequence when generating each part of the output. It significantly improves performance on long sequences.

Correct Answer: B - Extractive and abstractive

Text summarization is categorized as extractive (selecting important sentences from original text) and abstractive (generating new sentences that convey key information). Abstractive is more challenging.