Natural Language Processing (NLP) Certification Exam — Questions and Answers

Question 1: What distinguishes a Natural Language Processing certified professional from a non-certified practitioner?

Certified professionals exclusively work in larger organizations
There is no meaningful difference in competency
Certification validates competency through standardized assessment against established benchmarks (Correct answer)
Certified professionals always have more years of experience

Correct answer: Certification validates competency through standardized assessment against established benchmarks

Certification provides objective validation of competency through standardized assessment. While non-certified practitioners may be skilled, certification offers verified evidence that a professional meets established benchmarks for knowledge and performance.

Question 2: What is the MOST effective way for new NLP professionals to build competency in their field?

Combining formal education, mentored practice, and ongoing professional development (Correct answer)
Studying certification materials exclusively
Learning entirely through trial and error
Focusing solely on the most advanced topics

Correct answer: Combining formal education, mentored practice, and ongoing professional development

Building professional competency requires a multi-faceted approach: formal education provides foundational knowledge, mentored practice develops applied skills under guidance, and ongoing professional development ensures continuous growth and currency in the field.

Question 3: The proper order for Building Rapport is as follows:

Attention, Trust, Comfort
Attention, Comfort, Trust (Correct answer)
Comfort, Trust, Attention
Trust, Comfort, Attention

Correct answer: Attention, Comfort, Trust

Building rapport in NLP typically follows a progression that starts with gaining someone's attention. Once attention is established, the next step is to create a sense of comfort and ease, often through matching and mirroring. Finally, as comfort grows, trust can be built, forming a strong foundation for effective communication and influence.

Question 4: What distinguishes a Natural Language Processing certified professional from a non-certified practitioner?

Certified professionals always have more years of experience
There is no meaningful difference in competency
Certified professionals exclusively work in larger organizations
Certification validates competency through standardized assessment against established benchmarks (Correct answer)

Correct answer: Certification validates competency through standardized assessment against established benchmarks

Question 5: In the transformer encoder-decoder architecture for MT, what is the role of cross-attention?

Computes attention between source and target vocabularies
Normalizes attention scores across heads
Allows the decoder to attend to encoder outputs (Correct answer)
Allows the encoder to attend to itself

Correct answer: Allows the decoder to attend to encoder outputs

Cross-attention in the decoder lets each target position attend over all encoder hidden states to gather source context.

Question 6: What is 'domain adaptation' when applied to word embeddings?

Changing the embedding dimension to match a target model architecture
Translating embeddings from one language to another using a bilingual dictionary
Normalizing embedding magnitudes to the unit sphere
Further training or fine-tuning general embeddings on in-domain text to capture domain-specific vocabulary and meaning (Correct answer)

Correct answer: Further training or fine-tuning general embeddings on in-domain text to capture domain-specific vocabulary and meaning

General-purpose embeddings may not capture specialized terminology well, so continuing training on domain text (e.g., biomedical papers) adapts them to the target domain.

Question 7: What is the value of continuing education in sentiment analysis for NLP professionals?

It is only needed for recertification
It is primarily a social activity
It replaces workplace experience
It keeps professionals current with evolving standards and practices (Correct answer)

Correct answer: It keeps professionals current with evolving standards and practices

Continuing education ensures professionals stay current with the latest developments, standards, and best practices in their field.

Question 8: Which approach best demonstrates mastery of machine translation in NLP practice?

Applying principles to novel situations with sound judgment (Correct answer)
Avoiding complex scenarios
Relying entirely on technology
Following procedures without understanding

Correct answer: Applying principles to novel situations with sound judgment

True mastery involves understanding underlying principles well enough to apply them to new and unfamiliar situations with professional judgment.

Question 9: What is a 'sentence embedding,' and how does it differ from averaging individual word embeddings?

Sentence embeddings concatenate all word vectors into a single long vector
Sentence embeddings are obtained by taking the maximum word vector component-wise
Sentence embeddings are identical to averaged word embeddings but stored as sparse vectors
A sentence embedding is a single vector representing the whole sentence's meaning, often capturing word order and composition that simple averaging ignores (Correct answer)

Correct answer: A sentence embedding is a single vector representing the whole sentence's meaning, often capturing word order and composition that simple averaging ignores

Models like Sentence-BERT produce sentence embeddings via fine-tuned transformers that encode word order and inter-word relationships, unlike naive averaging which is order-invariant.

Question 10: What distinguishes a Natural Language Processing certified professional from a non-certified practitioner?

Certified professionals always have more years of experience
There is no meaningful difference in competency
Certified professionals exclusively work in larger organizations
Certification validates competency through standardized assessment against established benchmarks (Correct answer)

Correct answer: Certification validates competency through standardized assessment against established benchmarks

Question 11: Which evaluation approach asks bilingual judges to rate MT output on adequacy and fluency separately?

Direct Assessment (DA) on a 0-100 scale
BLEU with multiple references
Automatic post-editing evaluation
Human evaluation using adequacy/fluency scales (Correct answer)

Correct answer: Human evaluation using adequacy/fluency scales

The classic human MT evaluation protocol uses separate adequacy (meaning preserved?) and fluency (is target text natural?) rating scales judged by bilingual assessors.

Question 12: In multilingual NMT, what is the 'language token' prepended to the source sentence used for?

Signaling the desired target language to a single shared model (Correct answer)
Marking sentence boundaries for the tokenizer
Weighting the loss function per language
Indicating the domain of the text

Correct answer: Signaling the desired target language to a single shared model

A target-language tag (e.g., <2fr>) prepended to the input tells a universal NMT model which language to generate.

Question 13: Which foundational principle is MOST important for success in the Natural Language Processing profession?

Maintaining the minimum requirements for certification
Specializing in only one narrow area of practice
Commitment to continuous learning, ethical practice, and quality outcomes (Correct answer)
Maximizing financial returns on every engagement

Correct answer: Commitment to continuous learning, ethical practice, and quality outcomes

Success in any professional field requires a commitment to continuous learning to stay current, ethical practice to maintain trust and integrity, and a focus on quality outcomes that serve stakeholders and the public interest.

Question 14: In NLP practice, what is the best approach to quality improvement in machine translation?

Wait for problems to occur before acting
Copy what other organizations do without analysis
Use data-driven methods with measurable outcomes (Correct answer)
Make changes without measuring results

Correct answer: Use data-driven methods with measurable outcomes

Data-driven quality improvement with measurable outcomes ensures that changes actually produce the intended improvements and can be verified.

Question 15: In NLP practice, what is the best approach to quality improvement in sentiment analysis?

Use data-driven methods with measurable outcomes (Correct answer)
Copy what other organizations do without analysis
Wait for problems to occur before acting
Make changes without measuring results

Correct answer: Use data-driven methods with measurable outcomes

Data-driven quality improvement with measurable outcomes ensures that changes actually produce the intended improvements and can be verified.

Question 16: What is the PRIMARY purpose of obtaining NLP certification in Natural Language Processing?

To guarantee employment in the field
To bypass educational requirements
To demonstrate verified competency and adherence to professional standards (Correct answer)
To satisfy a personal achievement goal

Correct answer: To demonstrate verified competency and adherence to professional standards

Professional certification demonstrates that an individual has met established competency standards through verified assessment. It provides assurance to employers, clients, and the public that the certified professional possesses the knowledge and skills required for competent practice.

Question 17: Which foundational principle is MOST important for success in the Natural Language Processing profession?

Maintaining the minimum requirements for certification
Commitment to continuous learning, ethical practice, and quality outcomes (Correct answer)
Maximizing financial returns on every engagement
Specializing in only one narrow area of practice

Correct answer: Commitment to continuous learning, ethical practice, and quality outcomes

Question 18: Which IBM Model introduced the concept of word fertility in MT alignment?

IBM Model 3 (Correct answer)
IBM Model 4
IBM Model 1
IBM Model 2

Correct answer: IBM Model 3

IBM Model 3 introduced fertility, the number of target words that a single source word generates.

Question 19: What does the term 'valence' refer to in sentiment analysis?

The frequency of sentiment words in a corpus
The grammatical structure of a sentence
The confidence score of a classifier
The positive or negative direction of an emotional expression (Correct answer)

Correct answer: The positive or negative direction of an emotional expression

Valence captures whether an expression is positive or negative, forming the core dimension of sentiment polarity.

Question 20: Which of the following best describes the 'bag-of-words' representation after text preprocessing?

A graph of co-occurring word pairs
A sequential model preserving word order
An unordered collection of word frequencies ignoring grammar (Correct answer)
A hierarchical tree of syntactic phrases

Correct answer: An unordered collection of word frequencies ignoring grammar

Bag-of-words represents text as an unordered set of word counts, discarding grammar and word order information.

Question 21: What entails doing a grammar check on the given sentence

Matching
Parsing
To start doing (Correct answer)

Correct answer: To start doing

This question appears to be a play on words or refers to a specific NLP context. In standard English, 'doing a grammar check' means to analyze the grammar of a sentence. However, in some NLP contexts, 'doing' can refer to taking action or initiating a process, implying the action of performing the check rather than the analytical process itself.

Question 22: Which of the following is a key characteristic of convolutional neural networks (CNNs) when applied to text classification?

They apply filters over local n-gram windows to capture local features (Correct answer)
They use positional encodings to represent word order
They model long-range word dependencies through recurrent connections
They classify text by comparing embeddings to learned prototype vectors

Correct answer: They apply filters over local n-gram windows to capture local features

CNNs for text use 1D convolutional filters that slide over windows of consecutive words, learning to detect locally informative n-gram patterns useful for classification.

Question 23: Which professional attribute is most valued in machine translation within the NLP field?

Accountability and commitment to standards (Correct answer)
Prioritizing personal convenience
Working in isolation
Avoiding challenging situations

Correct answer: Accountability and commitment to standards

Accountability and commitment to professional standards build trust and ensure consistent, high-quality practice.

Question 24: What is the effect of choosing a larger vocabulary size in subword tokenization?

No effect on sequence length or vocabulary coverage
Shorter token sequences and fewer unique token types
Longer token sequences and fewer unique token types
Shorter token sequences but more unique token types to learn embeddings for (Correct answer)

Correct answer: Shorter token sequences but more unique token types to learn embeddings for

A larger vocabulary allows more complete words and longer subwords, reducing sequence length but requiring the model to learn more embedding vectors.

Question 25: What is the MOST effective way for new NLP professionals to build competency in their field?

Learning entirely through trial and error
Studying certification materials exclusively
Combining formal education, mentored practice, and ongoing professional development (Correct answer)
Focusing solely on the most advanced topics

Correct answer: Combining formal education, mentored practice, and ongoing professional development

Question 26: What is 'quantization' in the deployment of large language models?

Splitting the model across multiple GPUs to enable parallel decoding
Pruning the model by removing attention heads with low average activation
Reducing the sequence length of inputs to lower memory footprint
Representing model weights with lower-precision numbers (e.g., INT8 instead of FP32) to reduce size and speed up inference (Correct answer)

Correct answer: Representing model weights with lower-precision numbers (e.g., INT8 instead of FP32) to reduce size and speed up inference

Quantization compresses model weights by using fewer bits per parameter, significantly reducing memory and improving inference speed with minimal accuracy loss.

Question 27: What is the 'semantic orientation' of a word in sentiment analysis?

The grammatical case assigned to a noun
The part of speech tag assigned to a word
The degree to which a word is associated with positive or negative sentiment (Correct answer)
The topic category a word belongs to

Correct answer: The degree to which a word is associated with positive or negative sentiment

Semantic orientation quantifies how strongly a word skews positive or negative based on its associations and usage patterns.

Question 28: Which statement BEST describes the relationship between Natural Language Processing certification requirements and industry evolution?

Changes only occur when government mandates new requirements
Requirements evolve periodically to reflect advances in knowledge, technology, and practice standards (Correct answer)
Certification requirements never change once established
Requirements become less stringent over time

Correct answer: Requirements evolve periodically to reflect advances in knowledge, technology, and practice standards

Certification requirements evolve to keep pace with advances in professional knowledge, technological developments, and changes in practice standards. This ensures that certified professionals remain current and competent in a changing professional landscape.

Question 29: What is 'cross-lingual NER'?

Using multiple NER models in an ensemble
Training on one language and transferring the model to recognize entities in another (Correct answer)
NER that merges entities across parallel corpora
NER that switches languages mid-sentence

Correct answer: Training on one language and transferring the model to recognize entities in another

Cross-lingual NER trains on a high-resource language and transfers to a low-resource target language, leveraging multilingual embeddings like mBERT or XLM-R.

Question 30: Which statement BEST describes the relationship between Natural Language Processing certification requirements and industry evolution?

Requirements evolve periodically to reflect advances in knowledge, technology, and practice standards (Correct answer)
Certification requirements never change once established
Requirements become less stringent over time
Changes only occur when government mandates new requirements

Correct answer: Requirements evolve periodically to reflect advances in knowledge, technology, and practice standards

Question 31: What is the 'Pointwise Mutual Information' (PMI) method used for in sentiment analysis?

Evaluating the precision of sentiment classifiers
Measuring sentence perplexity in language models
Computing similarity between two sentiment lexicons
Estimating semantic orientation of phrases by comparing co-occurrence with positive vs. negative seed words (Correct answer)

Correct answer: Estimating semantic orientation of phrases by comparing co-occurrence with positive vs. negative seed words

PMI measures how much more often a phrase co-occurs with positive seed words versus negative ones to estimate its sentiment orientation.

Question 32: What is 'nucleus sampling' (top-p sampling) in language model decoding?

Applying beam search with p beams simultaneously
Sampling only from the single most probable next token at each step
Sampling from the smallest set of tokens whose cumulative probability exceeds a threshold p (Correct answer)
Sampling uniformly from the top-p% of the vocabulary by frequency

Correct answer: Sampling from the smallest set of tokens whose cumulative probability exceeds a threshold p

Top-p sampling dynamically selects a minimal set of tokens whose probabilities sum to at least p, then samples from that set — balancing diversity and coherence.

Question 33: What is the role of 'length normalization' in beam search for MT?

Normalizing attention weights so they sum to one
Ensuring source and target sentences have the same number of tokens
Dividing the log-probability score by sentence length to avoid penalizing longer hypotheses (Correct answer)
Rescaling embeddings by the square root of model dimension

Correct answer: Dividing the log-probability score by sentence length to avoid penalizing longer hypotheses

Without length normalization, beam search favors short translations because each additional token multiplies (reduces) the probability; dividing by length corrects this bias.

Question 34: What is the PRIMARY purpose of obtaining NLP certification in Natural Language Processing?

To satisfy a personal achievement goal
To guarantee employment in the field
To bypass educational requirements
To demonstrate verified competency and adherence to professional standards (Correct answer)

Correct answer: To demonstrate verified competency and adherence to professional standards

Question 35: What is 'back-translation' used for in low-resource neural machine translation?

Validating translations by re-translating to the source
Generating synthetic source sentences from target monolingual data (Correct answer)
Reversing the encoder-decoder order during inference
Translating the model weights back to source language

Correct answer: Generating synthetic source sentences from target monolingual data

Back-translation uses a reverse MT system to translate target-language monolingual text into the source, creating synthetic parallel data for training.

Question 36: Which component of a pipeline-based information extraction system comes directly AFTER NER?

Tokenization
Stopword removal
Relation extraction (Correct answer)
Sentence splitting

Correct answer: Relation extraction

Relation extraction identifies semantic relationships between entity pairs that NER has already identified, making it the natural downstream task.

Question 37: Which approach to MT does NOT require any parallel bilingual data during training?

Unsupervised MT using monolingual corpora only (Correct answer)
Supervised NMT
Transfer learning from multilingual models
Phrase-based SMT with phrase tables

Correct answer: Unsupervised MT using monolingual corpora only

Unsupervised MT methods (e.g., using denoising autoencoders and back-translation on monolingual data) require no parallel sentences.

Question 38: Which evaluation method tests word embeddings by checking whether cosine similarity rankings match human-rated similarity scores?

Intrinsic evaluation via word similarity benchmarks (Correct answer)
Perplexity measurement on a held-out corpus
Extrinsic evaluation on downstream NLP tasks
Cross-entropy scoring against a language model

Correct answer: Intrinsic evaluation via word similarity benchmarks

Intrinsic evaluation compares embedding-derived similarity rankings against human-annotated datasets like WordSim-353.

Question 39: What is the primary limitation of lexicon-based sentiment analysis?

It struggles with domain-specific language and context-dependent word meanings (Correct answer)
It is computationally too expensive for real-time use
It cannot handle binary (positive/negative) classification
It requires large amounts of labeled training data

Correct answer: It struggles with domain-specific language and context-dependent word meanings

Lexicon-based methods rely on fixed word-sentiment mappings that fail when domain context shifts the meaning (e.g., 'sick' meaning 'cool' in slang).

Question 40: Which foundational principle is MOST important for success in the Natural Language Processing profession?

Maintaining the minimum requirements for certification
Commitment to continuous learning, ethical practice, and quality outcomes (Correct answer)
Specializing in only one narrow area of practice
Maximizing financial returns on every engagement

Correct answer: Commitment to continuous learning, ethical practice, and quality outcomes

Question 41: What is the MOST effective way for new NLP professionals to build competency in their field?

Focusing solely on the most advanced topics
Learning entirely through trial and error
Combining formal education, mentored practice, and ongoing professional development (Correct answer)
Studying certification materials exclusively

Correct answer: Combining formal education, mentored practice, and ongoing professional development

Question 42: Which technique helps capture long-range dependencies in sentiment classification using deep learning?

Bag-of-words model
TF-IDF vectorization
LSTM or Transformer-based model (Correct answer)
Naive Bayes classifier

Correct answer: LSTM or Transformer-based model

LSTMs and Transformers maintain context across long sequences, capturing dependencies that bag-of-words models miss.

Question 43: Which of the following is a key advantage of character-level tokenization over word-level tokenization?

Shorter sequence lengths for the same text
Better capture of semantic meaning per token
No out-of-vocabulary (OOV) problem since any text can be represented (Correct answer)
Smaller model input sizes

Correct answer: No out-of-vocabulary (OOV) problem since any text can be represented

Character-level tokenization can represent any string using a small fixed alphabet, completely eliminating OOV issues.

Question 44: How does the NLP body of knowledge relate to daily professional practice?

It is relevant only for academic research
It is theoretical and has limited practical application
It provides the foundational framework that guides decision-making and standard practices (Correct answer)
It only applies during certification exams

Correct answer: It provides the foundational framework that guides decision-making and standard practices

The body of knowledge provides the foundational framework of principles, standards, and best practices that professionals use to guide their daily decision-making, ensure consistent quality, and maintain alignment with industry standards.

Question 45: What does 'document-level MT' aim to improve over sentence-level MT?

Vocabulary coverage across domains
BLEU scores on individual sentences
Coherence, coreference resolution, and consistency across sentences (Correct answer)
Training speed by batching entire documents

Correct answer: Coherence, coreference resolution, and consistency across sentences

Document-level MT models consider inter-sentence context to correctly resolve pronouns, maintain consistent terminology, and improve discourse coherence.

Question 46: The presumptions or beliefs of Neuro-linguistic programming (NLP) are

We process all Information through our senses
Modeling successful performance leads to excellence
Mind and body form a system. They are different expressions of the one person.
All of the above (Correct answer)

Correct answer: All of the above

All the listed options are core presuppositions of NLP. 'Modeling successful performance leads to excellence' is a foundational premise, 'Mind and body form a system' emphasizes their interconnectedness, and 'We process all Information through our senses' highlights the role of sensory input in our experience of reality.

Question 47: What is the 'exposure bias' problem in sequence-to-sequence training?

The embedding layer is exposed to raw bytes instead of subword tokens
At test time the model sees its own predictions, but during training it always sees ground-truth tokens (Correct answer)
Attention weights become biased toward early tokens in long sequences
The model is exposed to too many training examples, causing overfitting

Correct answer: At test time the model sees its own predictions, but during training it always sees ground-truth tokens

Exposure bias arises because teacher-forcing at train time hides prediction errors, making the model fragile to its own mistakes at inference.

Question 48: What is the MOST effective way for new NLP professionals to build competency in their field?

Focusing solely on the most advanced topics
Learning entirely through trial and error
Combining formal education, mentored practice, and ongoing professional development (Correct answer)
Studying certification materials exclusively

Correct answer: Combining formal education, mentored practice, and ongoing professional development

Question 49: In multi-class sentiment analysis, what does 'fine-grained' classification refer to?

Analyzing only short tweets rather than long reviews
Distinguishing between multiple languages
Performing sentiment analysis at the word level
Classifying sentiment into more than three categories such as very positive, positive, neutral, negative, very negative (Correct answer)

Correct answer: Classifying sentiment into more than three categories such as very positive, positive, neutral, negative, very negative

Fine-grained sentiment goes beyond positive/negative/neutral to capture a spectrum of intensity levels.

Question 50: What is this state known as when you withdraw yourself from an experience and observe it from a distance rather than being involved in it?

Imagination
Being disassociated (Correct answer)
Being disaffiliated
Being associated

Correct answer: Being disassociated

Being disassociated in NLP refers to the state of observing an experience from an external, objective perspective, as if watching a movie of oneself. This technique is often used to gain new insights, reduce emotional intensity, or reframe challenging situations by creating emotional distance from the event.

Question 51: Which component of a neural MT system is responsible for generating a fixed-length context vector in older encoder-decoder architectures (pre-attention)?

Embedding matrix
Positional encoding
Softmax layer
Final encoder hidden state (Correct answer)

Correct answer: Final encoder hidden state

In early seq2seq models, the last encoder hidden state compressed the entire source sentence into a single context vector passed to the decoder.

Question 52: Which approach is used in 'distant supervision' for sentiment analysis?

Using human-annotated labels for training
Training on synthetic data generated by a language model
Transferring labels from a related language
Automatically labeling data using heuristics such as review star ratings (Correct answer)

Correct answer: Automatically labeling data using heuristics such as review star ratings

Distant supervision exploits noisy but abundant labels (like star ratings) to train sentiment classifiers without manual annotation.

Question 53: What does 'zero-shot text classification' mean in the context of modern NLP?

Classifying text into categories that were not seen during model training (Correct answer)
Using zero regularization to prevent overfitting in classifiers
Classifying text with 100% accuracy without any errors
Training a model with no labeled data using only rule-based methods

Correct answer: Classifying text into categories that were not seen during model training

Zero-shot classification leverages large pre-trained language models to assign labels to new, unseen categories by understanding the semantic meaning of the label names.

Question 54: How does the NLP body of knowledge relate to daily professional practice?

It only applies during certification exams
It provides the foundational framework that guides decision-making and standard practices (Correct answer)
It is theoretical and has limited practical application
It is relevant only for academic research

Correct answer: It provides the foundational framework that guides decision-making and standard practices

Question 55: Which property of the transformer makes it more parallelizable during training compared to RNN-based MT models?

Self-attention operates on all positions simultaneously rather than sequentially (Correct answer)
Use of convolutional layers instead of recurrent connections
The encoder processes tokens from right to left
Positional encodings replace word embeddings

Correct answer: Self-attention operates on all positions simultaneously rather than sequentially

Transformer self-attention computes relationships between all token pairs in parallel, whereas RNNs must process tokens one at a time sequentially.

Question 56: How does the NLP body of knowledge relate to daily professional practice?

It is theoretical and has limited practical application
It only applies during certification exams
It is relevant only for academic research
It provides the foundational framework that guides decision-making and standard practices (Correct answer)

Correct answer: It provides the foundational framework that guides decision-making and standard practices

Question 57: Which statement BEST describes the relationship between Natural Language Processing certification requirements and industry evolution?

Certification requirements never change once established
Changes only occur when government mandates new requirements
Requirements become less stringent over time
Requirements evolve periodically to reflect advances in knowledge, technology, and practice standards (Correct answer)

Correct answer: Requirements evolve periodically to reflect advances in knowledge, technology, and practice standards

Question 58: What is a 'tokenizer mismatch' and why is it a critical issue in NLP deployments?

Applying tokenization to audio or image inputs instead of text
Running the tokenizer on a different operating system than it was built on
Using a different tokenizer at inference time than was used during model training, causing the model to receive unexpected token ID sequences (Correct answer)
Using two different tokenization libraries that produce the same output

Correct answer: Using a different tokenizer at inference time than was used during model training, causing the model to receive unexpected token ID sequences

A tokenizer mismatch means the model receives token IDs that don't correspond to the embeddings it learned, causing degraded or nonsensical outputs.

Question 59: What is the PRIMARY purpose of obtaining NLP certification in Natural Language Processing?

To guarantee employment in the field
To bypass educational requirements
To satisfy a personal achievement goal
To demonstrate verified competency and adherence to professional standards (Correct answer)

Correct answer: To demonstrate verified competency and adherence to professional standards

Question 60: What does 'opinion mining' typically extract beyond overall sentiment polarity?

The author's identity and publication date
Opinion holders, opinion targets, and sentiment expressions (Correct answer)
Document similarity scores
Grammar errors and spelling mistakes

Correct answer: Opinion holders, opinion targets, and sentiment expressions

Opinion mining identifies who holds the opinion, what the opinion is about, and the expressed sentiment toward it.

Question 61: What is the PRIMARY purpose of obtaining NLP certification in Natural Language Processing?

To demonstrate verified competency and adherence to professional standards (Correct answer)
To satisfy a personal achievement goal
To bypass educational requirements
To guarantee employment in the field

Correct answer: To demonstrate verified competency and adherence to professional standards

Question 62: In statistical machine translation, what does an 'alignment model' estimate?

The number of phrase pairs in the translation table
The correspondence between source and target words (Correct answer)
The probability of a target sentence given source
The language model score for the target sentence

Correct answer: The correspondence between source and target words

An alignment model (e.g., IBM Models) estimates the probability that source word i generated target word j.

Question 63: What does 'token alignment' refer to when tokenizing text for tasks like Named Entity Recognition (NER)?

Ensuring all tokens have the same vector length
Sorting tokens by their frequency in the corpus
Mapping subword tokens back to their original word boundaries to correctly assign labels (Correct answer)
Aligning the vocabulary of two different tokenizers

Correct answer: Mapping subword tokens back to their original word boundaries to correctly assign labels

In NER, word-level labels must be aligned to subword tokens; typically only the first subword of each word receives the label while others get a special ignore label.

Question 64: Which technique allows word embeddings from two different languages to be mapped into a shared cross-lingual vector space?

Learning a linear transformation (rotation) using bilingual anchor word pairs (Correct answer)
Concatenating the two monolingual embedding matrices
Applying PCA to both embedding sets independently
Re-training both language models jointly on a parallel corpus from scratch

Correct answer: Learning a linear transformation (rotation) using bilingual anchor word pairs

Cross-lingual alignment methods like VecMap find a rotation matrix that maps monolingual embeddings into a shared space using bilingual seed lexicons.

Question 65: What does 'zero-shot translation' mean in a multilingual NMT system?

Achieving BLEU=0 on a test set
Translating between a language pair never seen together in training (Correct answer)
Running inference without a decoder
Translating with zero training examples by few-shot prompting

Correct answer: Translating between a language pair never seen together in training

Zero-shot translation is the ability to translate between two languages that were never paired together in training data.

Natural Language Processing (NLP) Certification Exam

The NLP Certification Exam assesses proficiency in natural language processing concepts, techniques, and real-world applications including text analysis, machine translation, sentiment analysis, and foundational NLP methodologies.

Questions

90 min

Time Limit

70%

To Pass

Checking for saved progress...

Exam Rules

You can skip questions and return to them later
Flag questions for review before submitting
No feedback shown until you submit the entire exam
Unanswered questions count as wrong — answer everything
10 pretest questions are mixed in and don't affect your score
Timer auto-submits when time runs out
Your progress is auto-saved every 30 seconds