WorksheetsQuiz on Contextual Embeddings and Parsing
Total questions: 25
Worksheet time: 13mins
Which of the following is a key advantage of contextual embeddings compared to traditional word embeddings?
Fixed word meaning
Handles polysemy and context-dependent meaning
Lower computational cost
Rule-based representation
Which model introduced contextual embeddings using bidirectional attention?
Word2Vec
GloVe
BERT
FastText
In contextual embeddings, the meaning of a word depends on:
Only its frequency
Neighboring words (context)
Morphological structure alone
POS tagging only
Which of the following embeddings is static and not contextual?
ELMo
Word2Vec
BERT
GPT
Contextual embeddings are mainly generated using:
Decision trees
Neural networks with attention
Regular expressions
Rule-based parsers
In sequence-to-sequence models for machine translation, the encoder is responsible for:
Generating output text
Compressing input sequence into a representation
Tokenizing words
Evaluating accuracy
Which mechanism helps encoder-decoder models handle long sentences better?
Dropout
Beam Search
Attention
Stemming
The decoder in an encoder-decoder model mainly performs:
Sentence embedding
Target sequence generation
POS tagging
Grammar correction
Which neural network type was first widely used in machine translation encoder-decoder models?
CNN
RNN
Decision tree
K-means clustering
In Transformer-based translation models, parallelization is possible because:
They ignore grammar
Self-attention removes sequential dependency
They use statistical rules
They store all translations in memory
Constituency grammar represents sentences in terms of:
Word embeddings
Phrase structure
Dependency edges
TF-IDF
In constituency parsing, the sentence is represented as:
Dependency graph
Parse tree
Transition diagram
Hidden Markov model
Which of the following is an example of a constituent?
Only verbs
Only nouns
Noun phrases, verb phrases
Roots only
A major limitation of constituency parsing is:
It cannot represent sentence hierarchy
Ambiguity in phrase structures
It ignores syntax
It does not support recursion
In constituency grammars, 'The boy plays' can be split into:
Verb → plays, Object → the boy
NP → the boy, VP → plays
N → the, VP → boy plays
Root → plays
Dependency parsing focuses on:
Word frequency
Direct relations between words
Phrase structure
Vector embeddings
In dependency grammar, the root of the sentence usually corresponds to:
Article
Main verb
Noun phrase
Adjective
Which parsing approach is easier for machine learning models to compute?
Constituency parsing
Dependency parsing
Semantic parsing
Statistical parsing
In the dependency relation 'dog → barks', the word 'barks' is:
Head
Dependent
Modifier
Rootless
Which algorithm is commonly used for transition-based dependency parsing?
Earley's algorithm
Shift-Reduce
Beam search
Forward-Backward
Which of the following models uses contextual embeddings and encoder-decoder architecture?
BERT
Transformer
Naive Bayes
HMM
Constituency parsing is more aligned with:
Phrase structure grammar
Dependency grammar
Statistical grammar
Morphological grammar
Dependency parsing is more efficient than constituency parsing for:
Semantic role labeling
Identifying phrase boundaries
POS tagging
Word tokenization
In neural MT, beam search is used for:
Embedding generation
Optimal sequence decoding
Training optimization
Grammar correction
Which of the following correctly matches the concept and its application?
Contextual embeddings → fixed meaning
Encoder-Decoder → machine translation
Constituency parsing → word frequency
Dependency parsing → phrase segmentation
