What is the ultimate purpose of mapping text to vectors in these models?

What is the ultimate purpose of mapping text to vectors in these models?

• A. To create hard-coded grammatical rules.
• B. To optimize for exact string match.
• C. To minimize the dimensionality of vectors.
• D. To capture semantic relationships and enable similarity computations.

Answer: (D)

Mapping text to vectors creates a numeric space where meaning is encoded in the positions and relationships between points, not in literal spellings. The goal is to capture semantic relationships so the model can judge how similar two texts are, find related ideas, and perform reasoning using mathematical operations on those vectors. This enables tasks like retrieving documents that express the same concept, grouping paraphrases together, or solving analogies by vector arithmetic (for example, certain directional shifts in the vector space correspond to relationships between words or concepts). Because the representation focuses on meaning and usage rather than exact wording, it generalizes across synonyms and paraphrases, which is why measuring similarity and computing semantic relations are central aims. In contrast, hard-coded rules or exact string matching rely on surface form and rigidity, and while dimensionality reduction can be useful, it’s not the primary purpose of creating these vector representations.

Mapping text to vectors creates a numeric space where meaning is encoded in the positions and relationships between points, not in literal spellings. The goal is to capture semantic relationships so the model can judge how similar two texts are, find related ideas, and perform reasoning using mathematical operations on those vectors. This enables tasks like retrieving documents that express the same concept, grouping paraphrases together, or solving analogies by vector arithmetic (for example, certain directional shifts in the vector space correspond to relationships between words or concepts). Because the representation focuses on meaning and usage rather than exact wording, it generalizes across synonyms and paraphrases, which is why measuring similarity and computing semantic relations are central aims. In contrast, hard-coded rules or exact string matching rely on surface form and rigidity, and while dimensionality reduction can be useful, it’s not the primary purpose of creating these vector representations.