AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

WaveFake: A Data Set to Facilitate Audio Deepfake Detection

Joel Frank, Lea Schönherr

NeurIPS 2021 · Label-efficient and robust learning › Recognition and long-tail · approach: Dense features and foundation priors · 44 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 leave-one-out training across multiple distributions instead of a single training distribution (ablation)GMM LFCC · LJSpeech and JSUTaEER: 0.062 -> 0.058Improved
Use RawNet2 trained on multiple distributions rather than a single distribution (ablation)RawNet2 · LJSpeech and JSUTaEER: 0.04Improved
Use LFCC instead of MFCC features (ablation)GMM · LJSpeech and JSUTaEER: PWG: 0.241; MB-MelGAN: 0.257; WaveGlow: 0.214Improved
Leave out FB-MelGAN from training (ablation)GMM LFCC leave-one-out · JSUTEERHurt
Evaluate RawNet2 under simulated phone recording conditions (ablation)RawNet2 out-of-distribution models · JSUT and TTSEER: TTS: 0.144 to 0.549; JSUT MB-MelGAN: 0.000 to 0.898Hurt
Evaluate GMM LFCC under simulated phone recording conditions (ablation)GMM LFCC out-of-distribution models · JSUT and TTSEER: highest EER: 0.003Improved

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