Sign in Start Learning

← Question Bank

Deep Learning & Generative AI

Natural Language Processing Interview Questions and Answers

Interview questions on Natural Language Processing Interview Questions and Answers.

7 questions

All subtopics RNNs, LSTMs & Language Models Density Estimation & Training Paradigms Word Embeddings TF-IDF & Cosine Similarity N-Gram Language Models

RNNs, LSTMs & Language Models

Q1. What is the primary motivation for RNNs and LSTMs in NLP? How do you apply dropout in RNNs?

Sign in to bookmark

Density Estimation & Training Paradigms

Q2. What is Density Estimation? Why is a Language Model considered a Density Estimator?

Sign in to bookmark

Density Estimation & Training Paradigms

Q3. Language models are often called unsupervised, but some argue their mechanism is self-supervised/supervised. What are your thoughts?

Sign in to bookmark

Word Embeddings

Q4. Why do we need Word Embeddings? Compare Count-Based vs Prediction-Based Embeddings.

Sign in to bookmark

TF-IDF & Cosine Similarity

Q5. Given 5 documents and Query Q: "The early bird gets the worm", how does TF-IDF rank document relevance?

Sign in to bookmark

N-Gram Language Models

Q6. Should you choose an N-Gram or Neural Language Model for a tiny dataset (~10,000 tokens)?

Sign in to bookmark

N-Gram Language Models

Q7. Does increasing context length $N$ in N-gram models always improve performance?

Sign in to bookmark