AI Paper Atlas

Research map / Video / Video generation

JavisDiT++: Unified Modeling and Optimization for Joint Audio-Video Generation

Kai Liu, Yanhao Zheng, Kai Wang, Shengqiong Wu, Rongjunchen Zhang, Jiebo Luo, Dimitrios Hatzinakos, Ziwei Liu, Hao Fei, Tat-Seng Chua

ICLR 2026 · Video › Video generation · approach: Video-to-audio synthesis · 16 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
use modality-specific MoE FFNs instead of shared DiT with LoRA (ablation)JavisDiT++ · Architecture ablationJavisScore: 0.098 -> 0.153Improved
use modality-specific MoE FFNs instead of shared DiT with full fine-tuning (ablation)JavisDiT++ · Architecture ablationFVD: 269.3 -> 221.3Improved
use modality-specific MoE FFNs instead of shared DiT with full fine-tuning (ablation)JavisDiT++ · Architecture ablationFAD: 5.66 -> 5.51Improved
add AV-DPO with modality-aware micro reward (ablation)JavisDiT++ · AV-DPO reward designFVD: 221.3 -> 198.5Improved
add AV-DPO with modality-aware micro reward (ablation)JavisDiT++ · AV-DPO reward designDeSync: 0.807 -> 0.776Improved
use average-micro (modality-agnostic) reward in DPO instead of none (ablation)JavisDiT++ · AV-DPO reward designDeSync: 0.807 -> 0.810No clear effect
add reward normalization in AV-DPO pair ranking (ablation)JavisDiT++ · AV-DPO reward designFVD: 210.0 -> 198.5Improved
include ground-truth samples in AV-DPO pairs (ablation)JavisDiT++ · AV-DPO reward designFVD: 234.7 -> 198.5Improved
use interleave+offset TA-RoPE instead of vanilla position IDs (ablation)JavisDiT++ · Positional encoding ablationDeSync: 0.918 -> 0.807Improved
use interleave+offset TA-RoPE instead of vanilla position IDs (ablation)JavisDiT++ · Positional encoding ablationFAD (AudioCaps): 6.03 -> 4.65Improved
use interpolation instead of vanilla positional encoding (ablation)JavisDiT++ · Positional encoding ablationFAD (AudioCaps): 6.03 -> 12.87Hurt
train on 330K medium-quality data instead of 120K high-quality data (ablation)JavisDiT++ · AV-SFT data compositionFVD: 239.9 -> 221.3Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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