← Search

Chaoren Wang

2 accepted papers

2026

SpeechJudge: Towards Human-Level Judgment for Speech Naturalness

ICLR 2026poster

Aligning large generative models with human feedback is a critical challenge. In speech synthesis, this is particularly pronounced due to the lack of a large-scale human preference dataset, which hinders the development of models that truly align with human perception. To address this, we introduce…

Cited by 0SourceScholar
2025

Advancing Zero-shot Text-to-Speech Intelligibility across Diverse Domains via Preference Alignment

ACL 2025long

Modern zero-shot text-to-speech (TTS) systems, despite using extensive pre-training, often struggle in challenging scenarios such as tongue twisters, repeated words, code-switching, and cross-lingual synthesis, leading to intelligibility issues. To address these limitations, this paper leverages pre…