AI Paper Atlas

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

From Natural Alignment to Conditional Controllability in Multimodal Dialogue

Zeyu Jin, Songtao Zhou, Haoyu Wang, Minghao Tian, Kaifeng Yun, Zhuo Chen, Xiaoyu Qin, Jia Jia

ICLR 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
fine-tune on MM-DiaHiggs-Audio-V2 · MM-Dia Test (Description control)WER: 31.3 -> 4.5Improved
fine-tune on MM-DiaHiggs-Audio-V2 · MM-Dia Test (Description control)cp-WER: 104.8 -> 33.8Improved
fine-tune on MM-DiaHiggs-Audio-V2 · MM-Dia Test (Description control)sa-SIM: 0.48 -> 0.45Hurt
fine-tune on MM-DiaHiggs-Audio-V2 · MM-Dia Test (Description control)Human-MOS quality: 3.580 -> 4.440Improved
fine-tune on MM-DiaDia · MM-Dia Test (Description control)WER: 19.991 -> 29.071Hurt
fine-tune on MM-DiaDia · MM-Dia Test (Description control)Human-MOS quality: 2.410 -> 2.890Improved
use cascaded GPT style prompts instead of end-to-end HarmoniVoxHiggs-Audio-SFT · MM-Dia-BenchWER: 21.223 -> 5.793Improved
provide style cues implicitly via visual context instead of explicit descriptionHiggs-Audio-SFT · MM-Dia-BenchGemini-as-Judge similarity: 4.640 -> 3.022Hurt
calibrate SRT subtitles (ablation)SRT subtitles · Subtitle timestamptimestamp accuracy: 0.179 -> 0.857Improved
use buffer b=2 instead of b=1 (ablation)Qwen 7B segmentation · Dialogue segmentationF1 overlap: 0.771 -> 0.866Improved
increase buffer from b=1 to b=5 (ablation)Qwen 72B segmentation · Dialogue segmentationF1 overlap: 0.947 -> 0.979Improved
use Gemini instead of Qwen for non-verbal sound annotationannotation · MM-Dia annotationcompleteness: 1.25 -> 4.66Improved

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