Embedding models trained on Census surname, first-name, and voter file data improve probabilistic race prediction for uncommon surnames, with full-name embeddings delivering the largest gains especially for Hispanic and Asian individuals.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Using Embedding Models to Improve Probabilistic Race Prediction
Embedding models trained on Census surname, first-name, and voter file data improve probabilistic race prediction for uncommon surnames, with full-name embeddings delivering the largest gains especially for Hispanic and Asian individuals.