AI Paper Atlas

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

Convex Low-resource Accent-Robust Language Detection in Speech Recognition

Miria Feng, William Tan, Mert Pilanci

ICML 2026 · Large language models › Language, safety and interpretability · approach: Speech and audio language models · 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
add CLD convex language-detection head on Whisper-Small instead of default detectorWhisper-Small · Multiclass ASR, 5 languagesWER: 139.37 -> 31.74Improved
add CLD head instead of default detectorMMS-1B · Multiclass language IDdetection accuracy: 0.6701 -> 0.9702Improved
use convex CLD head instead of non-convex NN head (ablation)Whisper-Small · Multiclass language IDdetection accuracy: 0.7737 -> 0.9715Improved
use CLD instead of NN head on Whisper-Large-V3 (ablation)Whisper-Large-V3 · Multiclass language IDdetection accuracy: 0.9605 -> 0.9806Improved
use CLD instead of linear or kernel SVM on MMS-1B (ablation)MMS-1B · Multiclass language IDdetection accuracy: 0.5701 (kernel SVM) -> 0.9702Improved
use KNN as language detector (ablation)Whisper-Small/Large/MMS · Multiclass language IDdetection accuracy: 0.7154 (default) -> 0.6123 (KNN) on Whisper-SmallHurt
use CLD instead of fine-tuned NN or WSP-SFT as training data shrinks (ablation)Whisper-Small · Binary English/Mandarin accents, 100-10000 samplesaccuracy and WER: CLD 96.94% to 99.14% accuracy across sizesImproved
use CLD instead of NN or WSP-SFT for training cost (ablation)Whisper-Small · Multiclass datasettraining time (s): 840.30 -> 64.45 (NN), 1,096.74 (WSP-SFT)Improved
use CLD instead of NN on dialect-imbalanced English/Chinese (ablation)Whisper-Small · English/Chinese dialects, 500 samplesaccuracy (total): 0.5580 -> 0.9695Improved
use CLD with Whisper-Large-V3 versus NN on WER (ablation)Whisper-Large-V3 · Multiclass ASRWER: 29.25 -> 28.60No 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