Core-set - NeurIPS 2025

Core-set

1 paper across 1 session

Poster Session 3

Thursday, December 4, 2025 · 11:00 AM → 2:00 PM

CoreaSpeech: Korean Speech Corpus via JAMO-based Coreset Selection for Efficient and Robust Korean Speech Generation

#1916 · Ki-Joong Kwon, Junho So, Sang-Hoon Lee

We introduce CoreaSpeech, a publicly released 700 h Korean TTS corpus built via a Korean-specific text normalizer and a novel jamo-based coreset selection pipeline.