Bird
0
0

Why does t-SNE sometimes produce different visualizations on multiple runs with the same data and parameters?

hard📝 Conceptual Q10 of 15
NLP - Word Embeddings
Why does t-SNE sometimes produce different visualizations on multiple runs with the same data and parameters?
ABecause t-SNE changes input data internally each run
BBecause t-SNE uses random initialization affecting the embedding layout
CBecause t-SNE output depends on the order of input samples
DBecause t-SNE automatically changes perplexity each run
Step-by-Step Solution
Solution:

  1. Step 1: Understand t-SNE randomness

    t-SNE starts with random initial positions, so results can vary unless random seed is fixed.

  2. Step 2: Check other options

    Input data and parameters do not change automatically; order does not affect output significantly.

  3. Final Answer:

    Because t-SNE uses random initialization affecting the embedding layout -> Option B

  4. Quick Check:

    Random init causes visualization differences [OK]
Quick Trick: Fix random_state to get consistent t-SNE plots [OK]
Common Mistakes:
MISTAKES
  • Thinking input changes
  • Assuming automatic parameter changes
  • Blaming input order

Want More Practice?

15+ quiz questions · All difficulty levels · Free

Free Signup - Practice All Questions
More NLP Quizzes
Sentiment Analysis Advanced - Fine-grained sentiment (5-class) - Quiz 7medium Sentiment Analysis Advanced - Sentiment with context (sarcasm, negation) - Quiz 6medium Sequence Models for NLP - Embedding layer usage - Quiz 9hard Sequence Models for NLP - Why sequence models understand word order - Quiz 1easy Sequence Models for NLP - RNN for text classification - Quiz 9hard Text Similarity and Search - Jaccard similarity - Quiz 4medium Topic Modeling - Why topic modeling discovers themes - Quiz 14medium Word Embeddings - Word similarity and analogies - Quiz 8hard Word Embeddings - Pre-trained embedding usage - Quiz 15hard Word Embeddings - Pre-trained embedding usage - Quiz 9hard