一作NeurIPS 2026Poster
Is Text All You Need? Text as a Universal Information Bottleneck for Speech LLMs
Ming-Hao Hsu · Yuxuan Hu · Shujie LIU · Jinyu Li · Zhizheng Wu
多模态OpenReview
首次获取该候选人的完整档案及 StartLux 评估共 20 积分；每位用户单独解锁，之后查看免费。主动刷新 20 积分。
实时查找电子邮箱，费用不超过20积分；未找到邮箱不扣费。
Ming-Hao Hsu · Yuxuan Hu · Shujie LIU · Jinyu Li · Zhizheng Wu
Ryan Soh-Eun Shim · Kwanghee Choi · Kalvin Chang · Ming-Hao Hsu · Florian Eichin · Zhizheng Wu · Alane Suhr · Michael A. Hedderich · David Harwath · David R. Mortensen · Barbara Plank
Multilingual speech foundation models such as Whisper are trained on web-scale data, where data for each language consists of a myriad of regional varieties. However, different regional varieties often employ different scripts to write the same language, rendering speech recognition output also s...