AI Paper Atlas

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

Zero-AVSR: Zero-Shot Audio-Visual Speech Recognition with LLMs by Learning Language-Agnostic Speech Representations

Jeong Hun Yeo, Minsu Kim, Chae Won Kim, Stavros Petridis, Yong Man Ro

ICCV 2025 · Large language models › Language, safety and interpretability · approach: Sign and cross-lingual translation · 13 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 MARC audio-visual data (same 8 languages) beyond MuAViC (ablation)Cascaded Zero-AVSR · MuAViC avg CER, unseen Russianavg CER (%): 48.3 -> 28.3Improved
increase training languages from 8 to 40 in MARC (ablation)Cascaded Zero-AVSR · Unseen Russianzero-shot CER (%): 61.0 -> 49.5Improved
increase training languages from 40 to 81 (ablation)Cascaded Zero-AVSR · Unseen Russianzero-shot CER (%): 49.5 -> 40.0Improved
increase training languages from 8 to 81 (ablation)Cascaded Zero-AVSR · Unseen Spanishzero-shot CER (%): 29.9 -> 19.0Improved
use larger/better de-romanizer LLM (Llama3.2-3B to GPT-4o-mini) (ablation)Cascaded Zero-AVSR · MuAViC avg of 8 languagesavg CER (%): 35.7 -> 19.5Improved
use Llama3.1-70B instead of Llama3.1-8B as de-romanizer (ablation)Cascaded Zero-AVSR · MuAViC avg of 8 languagesavg CER (%): 27.3 -> 21.3Improved
add Spanish data (same Romance family) to training (ablation)Cascaded Zero-AVSR · Unseen SpanishCER (%): 40.1 -> 11.5Improved
add Turkish and Korean (different family) data (ablation)Cascaded Zero-AVSR · Unseen Romance languages, FrenchCER (%): 55.8 -> 59.7Hurt
add Spanish then Italian data on top of baseline (ablation)Cascaded Zero-AVSR · Unseen PortugueseCER (%): 48.0 -> 41.9Improved
add Spanish data (ablation)Cascaded Zero-AVSR · Unseen FrenchCER (%): 55.8 -> 56.6No clear effect
fine-tune LLM with multi-task training and feed speech features directly (Zero-AVSR vs Cascaded) with same Llama3.2-3B (ablation)Zero-AVSR · Unseen Russianzero-shot CER (%): 79.8 -> 45.4Improved
use GPT-4o-mini for both romanizing and de-romanizing instead of Llama3.1-70B for both (ablation)Romanization pipeline · MuAViC reconstruction testavg CER: 13.8 -> 4.0Improved

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