PhD student, Cornell University
1 paper at NeurIPS 2025
We introduce the first method for translating text embeddings from one vector space to another without any paired data, encoders, or predefined sets of matches.