AI Paper Atlas

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

Unlocking Strong Supervision: A Data-Centric Study of General-Purpose Audio Pre-Training Methods

Xuanru Zhou, Yiwen Shao, Wei-Cheng Tseng, Dong Yu

CVPR 2026 · Vision-language and multimodal › Cross-modal learning · approach: Self-supervised audio-visual representation learning · 1 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 UTS tag supervision from 400k clips instead of AudioSet tags from 2M clipsMTC pre-training · VoxCeleb2 linear probe (mean-pooling)accuracy: 18.84 -> 37.10Improved
use UTS tag supervision from 400k clips instead of AudioSet tags from 2M clipsMTC pre-training · FSD-50k linear probe (mean-pooling)mAP: 0.656 -> 0.459Hurt
use UTS-based MTC instead of AudioSet MTC for music question answeringMTC pre-training · AIR-Bench Music QAGPT-4o score: 5.61 -> 6.16Improved
replace original captions with high-fidelity Qwen3-Omni-Captioner summariesCaptioning-scratch · AudioCaps and retrievalRecall (retrieval): 49.2 -> 50.2Improved
replace original captions with high-fidelity captionsContrastive-scratch · linear probing, attention poolingmixed across tasks: FSD-50k 0.534 -> 0.514; VoxCeleb2 70.18 -> 71.63Mixed
add the tag multi-tag classification loss to the captioning objective (multi-task) (ablation)Captioning-scratch · FSD-50k linear probe (mean-pooling)mAP: 0.439 -> 0.485Improved
use parallel decoding objective instead of MTC (ablation)UTS tag pre-training · AIR-Bench speech QAaccuracy: MTC gender 92.64 vs PAR 37.40Hurt
switch from mean-pooling to multi-head attention pooling (ablation)Contrastive-scratch (Ours) · VoxCeleb2 linear probeaccuracy: 33.88 -> 71.63Improved
increase UTS vocabulary size beyond 1.5k to 2k (ablation)MTC pre-training · linear probing and retrievalnormalized average performanceHurt
increase UTS vocabulary size (ablation)captioning · captioningnormalized average performanceNo clear effect

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