AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

DIFFA: Large Language Diffusion Models Can Listen and Understand

Jiaming Zhou, Hongjie Chen, Shiwan Zhao, Jian Kang, Jie Li, Enzhi Wang, Yujie Guo, Haoqin Sun, Hui Wang, Aobo Kong, Yong Qin, Xuelong Li

AAAI 2026 · Large language models › Language, safety and interpretability · approach: Speech and audio language models · 14 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 diffusion LLM (LLaDA) backbone instead of autoregressive LLaMA 3.1 (ablation)DIFFA (dual adapter, LLaMA3 data) · MMAUaccuracy: 28.40 -> 51.71Improved
use diffusion LLM (LLaDA) backbone instead of autoregressive LLaMA 3.1 (ablation)DIFFA (dual adapter, LLaMA3 data) · MMSUaccuracy: 38.40 -> 54.72Improved
add acoustic adapter (dual vs single adapter) (ablation)DIFFA · MMAUaccuracy: 47.70 -> 49.71Improved
add acoustic adapter (dual vs single adapter) (ablation)DIFFA · MMSUaccuracy: 52.88 -> 56.04Improved
use Qwen3-generated instruction data instead of LLaMA3 (ablation)DIFFA · VoiceBenchaccuracy: 37.17 -> 48.22Improved
use Qwen3-generated instruction data instead of LLaMA3 (ablation)DIFFA · MMAUaccuracy: 51.71 -> 49.71Mixed
rewrite Qwen3 captions with LLaDA (ablation)DIFFA · Avg of MMAU/MMSU/VoiceBenchaccuracy: 51.32 -> 51.15No clear effect
use LLaDA-generated instruction data instead of LLaMA3 (ablation)DIFFA · Avg of MMAU/MMSU/VoiceBenchaccuracy: 47.86 -> 50.34Improved
train DIFFA with 127h synthetic data vs Qwen2-Audio large-scale dataDIFFA · MMSUaccuracy: 53.27 -> 56.04Improved
DIFFA vs Qwen2-Audio on VoiceBenchDIFFA · VoiceBench overallscore: 55.34 -> 48.22Hurt

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