Patent2015
Computing numeric representations of words in a high-dimensional space
Google LLC · USPTO — US9037464B1 · 2015
Abstract
Covers a system that maps words to points in a high-dimensional space using a neural network trained on surrounding context, such that words appearing in similar contexts end up near one another. The claims describe the training procedure and the use of the resulting vectors for comparing and retrieving words.
Why it matters
The word-embedding idea as granted intellectual property. Useful for seeing how a technique that became foundational infrastructure was framed in claim language — and a reminder that the paper and the patent are separate artefacts with different purposes.
patentretrievalarchitecture
Read the source
https://patents.google.com/patent/US9037464B1/en