An interactive online dictionary wants to add a widget that displays words used in similar contexts. A Machine Learning Specialist is asked to provide word features for the downstream nearest…

Question

Options

• ACreate one-hot word encoding vectors.
• BProduce a set of synonyms for every word using Amazon Mechanical Turk.
• CCreate word embedding vectors that store edit distance with every other word.
• DDownload word embeddings pre-trained on a large corpus.

• A
  3% (1)
• B
  6% (2)
• C
  14% (5)
• D
  78% (28)

Explanation

The widget needs to find words used in similar contexts - a semantic similarity task. Pre-trained word embeddings (e.g., Word2Vec, GloVe, FastText) are dense vector representations learned from large corpora that encode semantic and syntactic relationships; words used in similar contexts end up with similar vectors, making them ideal input features for a nearest neighbor model. One-hot vectors (A) are sparse and encode no semantic relationships. Mechanical Turk synonyms (B) are brittle, don't scale, and don't produce numeric features. Edit distance (C) measures character-level string differences, not contextual or semantic similarity.

No community discussion yet for this question.