🎯 Words as Vectors
Learn how words are transformed into numerical representations that capture meaning
Your Progress
0 / 5 completed←
Previous Module
Text Tokenization Playground
Words as Vectors
Word embeddings convert words into dense vectors of numbers, capturing semantic meaning and relationships. Words with similar meanings have similar vectors, enabling mathematical operations on language.
🎯 Why Word Embeddings?
✓
Semantic meaning: Capture word relationships
✓
Dimensionality: Dense 100-300D vectors
✓
Similarity: Measure word closeness
✓
Algebra: Perform vector arithmetic
🔢 Vector Representation
Word: "king"
Vector (300D): [0.25, -0.18, 0.41, 0.09, ...]
Word: "queen"
Vector (300D): [0.23, -0.16, 0.39, 0.11, ...]
Similar words have similar vector values, enabling semantic computations.
📏
Cosine Similarity
Measure angle between vectors (0-1 scale)
🧮
Vector Arithmetic
king - man + woman = queen
🗺️
Embedding Space
High-dimensional semantic landscape
💡 Famous Example
king - man + woman = ?
👑 - 👨 + 👩 = 👸
Result: queen