Dual Information Speech Language Models for Emotional Conversations
Chun Wang 1 3 , Chenyang Liu 1 , Wenze Xu 1 2 , Weihong Deng 1 3 · 2025
Read Dual Information Speech Language Models for Emotional Conversations by Chun Wang 1 3 , Chenyang Liu 1 , Wenze Xu 1 2 , Weihong Deng 1 3 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-08095.txt.
Description
Chun Wang, Chenyang Liu, Wenze Xu, and Weihong Deng present a speech language model for emotional conversation, in two thousand twenty-five. Text-only chat systems miss cues in the voice that carry feeling and intent. Speech models can hear those cues, but many are built by extending frozen text models and lose context in the process.
The team trace the problem to tangled information and poor training plans. They add two unlike adapters that split voice cues from word content. A weakly guided training plan keeps context strong while training only the adapters. Tests showed competitive results on emotional conversation tasks, with gains in both data and parameter efficiency.
Catalog record
| CREATOR | Chun Wang 1 3 , Chenyang Liu 1 , Wenze Xu 1 2 , Weihong Deng 1 3 |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-08095