> ML_ALGORITHM // CONTINUOUS-BAG-OF-WORDS-SKIPGRAM-WORD2VEC_v1.0
Word2Vec (CBOW & Skip-Gram)
Pioneering neural word embedding technique that maps vocabulary tokens into dense vector spaces capturing linear semantic and syntactic regularities.
Static Word Embeddingsself-supervisedhigh-intrinsiclarge (>100k)
Back to All AlgorithmsComputational Complexity
Training Complexity:O(epochs * corpus_size * window * negative_samples)
Inference Complexity:O(1) dictionary vector lookup
Hardware Profile
CPU Friendly:Yes
Requires GPU:No
Memory Footprint:low
Interpretability & Data
Interpretability Tier:high-intrinsic
Training Data Needs:large (>100k)
Interpretability Assessment
Embedding vector geometry satisfies linear semantic analogies (e.g., King - Man + Woman = Queen).
Suitable Tasks & Supported Modalities
Suitable Tasks:
feature extractiontoken embedding
Supported Modalities:
text
Implementing Libraries
Foundational Literature
Common Pitfalls & Warnings
- Conflating multiple meanings of polysemous words (e.g., "bank" of a river vs financial "bank") into a single static vector
- Zero vectors or crash on out-of-vocabulary (OOV) tokens