Cross-Lingual Word Embeddings
English ↔ Punjabi bilingual vector space alignment
- Aligned English and Punjabi word vector spaces using FastText embeddings and bilingual lexicon mapping, enabling semantic similarity comparison across two linguistically distinct languages
- Evaluated cross-lingual alignment quality with Precision@1 and Precision@5 on held-out bilingual lexicons
- Contributed open NLP resources for the low-resource Punjabi language