AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

Making Visual Dialogue More Engaging: A New Task, Method, and Metric

Guanghui Ye, Huan Zhao, Yingxue Gao, Zhixue Zhao, Kehan Wang, Xupeng Zha, Zhihua Jiang

AAAI 2026 · Vision-language and multimodal › Cross-modal learning · approach: Audio-visual speech separation · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add audio utterances (ablation)VITA-DM (7B) · Image-Chat_AVDCIDEr: 1.08 -> 1.28Improved
add image captions (ablation)VITA-DM (7B) · Image-Chat_AVDCIDEr: 1.03 -> 1.28Improved
add text utterances (ablation)VITA-DM (7B) · Image-Chat_AVDCIDEr: 0.86 -> 1.28Improved
add ICA pre-training task (ablation)VITA-DM (7B) · Image-Chat_AVDCIDEr: 0.98 -> 1.28Improved
add CCL pre-training task (ablation)VITA-DM (7B) · Image-Chat_AVDCIDEr: 1.09 -> 1.28Improved
add NUP pre-training task (ablation)VITA-DM (7B) · Image-Chat_AVDCIDEr: 1.20 -> 1.28Improved
represent emotion as audio instead of text label (ablation)VITA-DM (7B) · Image-Chat_AVDCIDEr: 1.10 -> 1.28Improved
represent emotion as audio instead of label, Photo-Chat (ablation)VITA-DM (7B) · Photo-Chat_AVDCIDEr: 1.09 -> 1.24Improved
use sentiment-aware Parler-TTS instead of UniCATS (ablation)VITA-DM (7B) · Image-Chat_AVDCIDEr: 1.14 -> 1.28Improved
train with VITA-DM on GPT-2-MediumGPT-2-M-345M · Image-Chat_AVDBLEU-1: 8.01 -> 10.34Improved
train with VITA-DM on Llama-2-7BLlama-2-7B · Image-Chat_AVDCIDEr: 1.05 -> 1.28Improved
add audio pre-training tasks but remove all pre-training tasks, audio vs label (ablation)VITA-DM · Image-Chat_AVDCIDErMixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy