Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chen, Tuochao, Veluri, Bandhav, Gong, Hongyu, Gollakota, Shyamnath
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2511.11124
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917080146640896
author Chen, Tuochao
Veluri, Bandhav
Gong, Hongyu
Gollakota, Shyamnath
author_facet Chen, Tuochao
Veluri, Bandhav
Gong, Hongyu
Gollakota, Shyamnath
contents Dialogue models falter in noisy, multi-speaker environments, often producing irrelevant responses and awkward turn-taking. We present AV-Dialog, the first multimodal dialog framework that uses both audio and visual cues to track the target speaker, predict turn-taking, and generate coherent responses. By combining acoustic tokenization with multi-task, multi-stage training on monadic, synthetic, and real audio-visual dialogue datasets, AV-Dialog achieves robust streaming transcription, semantically grounded turn-boundary detection and accurate responses, resulting in a natural conversational flow. Experiments show that AV-Dialog outperforms audio-only models under interference, reducing transcription errors, improving turn-taking prediction, and enhancing human-rated dialogue quality. These results highlight the power of seeing as well as hearing for speaker-aware interaction, paving the way for {spoken} dialogue agents that perform {robustly} in real-world, noisy environments.
format Preprint
id arxiv_https___arxiv_org_abs_2511_11124
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AV-Dialog: Spoken Dialogue Models with Audio-Visual Input
Chen, Tuochao
Veluri, Bandhav
Gong, Hongyu
Gollakota, Shyamnath
Computation and Language
Artificial Intelligence
Computer Vision and Pattern Recognition
Multimedia
Sound
Dialogue models falter in noisy, multi-speaker environments, often producing irrelevant responses and awkward turn-taking. We present AV-Dialog, the first multimodal dialog framework that uses both audio and visual cues to track the target speaker, predict turn-taking, and generate coherent responses. By combining acoustic tokenization with multi-task, multi-stage training on monadic, synthetic, and real audio-visual dialogue datasets, AV-Dialog achieves robust streaming transcription, semantically grounded turn-boundary detection and accurate responses, resulting in a natural conversational flow. Experiments show that AV-Dialog outperforms audio-only models under interference, reducing transcription errors, improving turn-taking prediction, and enhancing human-rated dialogue quality. These results highlight the power of seeing as well as hearing for speaker-aware interaction, paving the way for {spoken} dialogue agents that perform {robustly} in real-world, noisy environments.
title AV-Dialog: Spoken Dialogue Models with Audio-Visual Input
topic Computation and Language
Artificial Intelligence
Computer Vision and Pattern Recognition
Multimedia
Sound
url https://arxiv.org/abs/2511.11124