Paper: DOI badge, link and citation
Browse files
README.md
CHANGED
|
@@ -16,6 +16,10 @@ tags:
|
|
| 16 |
|
| 17 |
# Cue v5: turn-taking for voice agents, trained on real conversation
|
| 18 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
Cue listens to a call and decides, every 160 ms, what a voice agent should do: keep talking,
|
| 20 |
**PAUSE** or **STOP** while it speaks; **RESPOND** or keep waiting after the caller stops.
|
| 21 |
v5 is Cue v4 retrained on real human conversation as well as synthetic calls, and it can hear
|
|
@@ -134,6 +138,20 @@ systems on the same dev set include VAP (0.841 / 0.957) and Kyutai's semantic VA
|
|
| 134 |
- The default settings trade synthetic false stops for real-speech recall; tune per deployment.
|
| 135 |
- Indian-English and Hinglish backchannels ("haan", "achha") are only in the synthetic data.
|
| 136 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 137 |
## Licences
|
| 138 |
|
| 139 |
- `encoder/`: Whisper small's encoder, Apache-2.0 (position table shortened; see NOTICE).
|
|
|
|
| 16 |
|
| 17 |
# Cue v5: turn-taking for voice agents, trained on real conversation
|
| 18 |
|
| 19 |
+
[](https://doi.org/10.5281/zenodo.23144482)
|
| 20 |
+
|
| 21 |
+
**Paper:** *Cue: A Conversation Reflex Model for Barge-In and End-of-Turn Control in Cascaded Voice Agents*, preprint, 2026. [doi:10.5281/zenodo.23144482](https://doi.org/10.5281/zenodo.23144482) 路 [OpenAIRE research archive](https://explore.openaire.eu/search/result?pid=10.5281%2Fzenodo.23144482)
|
| 22 |
+
|
| 23 |
Cue listens to a call and decides, every 160 ms, what a voice agent should do: keep talking,
|
| 24 |
**PAUSE** or **STOP** while it speaks; **RESPOND** or keep waiting after the caller stops.
|
| 25 |
v5 is Cue v4 retrained on real human conversation as well as synthetic calls, and it can hear
|
|
|
|
| 138 |
- The default settings trade synthetic false stops for real-speech recall; tune per deployment.
|
| 139 |
- Indian-English and Hinglish backchannels ("haan", "achha") are only in the synthetic data.
|
| 140 |
|
| 141 |
+
## Citation
|
| 142 |
+
|
| 143 |
+
```bibtex
|
| 144 |
+
@misc{cue2026,
|
| 145 |
+
title = {Cue: A Conversation Reflex Model for Barge-In and End-of-Turn Control in Cascaded Voice Agents},
|
| 146 |
+
author = {Nishanth Tarun A, Joshua and Ajitesh Varun A, Joel},
|
| 147 |
+
year = {2026},
|
| 148 |
+
publisher = {Zenodo},
|
| 149 |
+
doi = {10.5281/zenodo.23144482},
|
| 150 |
+
url = {https://doi.org/10.5281/zenodo.23144482},
|
| 151 |
+
note = {Preprint}
|
| 152 |
+
}
|
| 153 |
+
```
|
| 154 |
+
|
| 155 |
## Licences
|
| 156 |
|
| 157 |
- `encoder/`: Whisper small's encoder, Apache-2.0 (position table shortened; see NOTICE).
|