Audio Deepfake Detection on In-the-wild
0.8EERXLS-R-1B
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| XLS-R-1BTraining Strategy=DOSS-Weight, #Params=965M, #Hours=12k2025.12 | 0.8 | — | — | — | |
| XLS-R-300MTraining Strategy=DOSS-Weight, #Params=317M, #Hours=12k2025.12 | 0.81 | — | — | — | |
| SOTATraining Strategy=Existing Benchmarks2025.12 | 1.23 | — | — | — | |
| XLS-R-2BTraining Strategy=Naive Aggregation, #Params=2.2B, #Hours=74k2025.12 | 1.23 | — | — | — | |
| XLS-R-1BTraining Strategy=Naive Aggregation, #Params=965M, #Hours=74k2025.12 | 1.37 | — | — | — | |
| MMS-1BTraining Strategy=Naive Aggregation, #Params=965M, #Hours=74k2025.12 | 1.82 | — | — | — | |
| AFSSVenue=Ours2026.03 | 2.7 | — | — | — | |
| MMS-300MTraining Strategy=Naive Aggregation, #Params=317M, #Hours=74k2025.12 | 2.9 | — | — | — | |
| SCLVenue=Interspeech ’242026.03 | 4.52 | — | — | — | |
| BreathNetPublication=Ours2026.02 | 4.7 | — | — | — | |
| MamBo-3-Hydra-N3Params (M)=319.372026.01 | 4.97 | — | — | — | |
| MamBo-4-Hydra-N1Params (M)=318.022026.01 | 5.17 | — | — | — | |
| Nes2Net-XPublication=TIFS 20252026.02 | 5.52 | — | — | — | |
| MamBo-1-Mamba2-N2Params (M)=317.482026.01 | 5.57 | — | — | — | |
| HyperPotterParams (M)=317.872026.02 | 5.72 | — | — | — | |
| Fake-Mamba (L)Params (M)=319.722026.01 | 5.85 | — | — | — | |
| LSR+LSAPublication=ICASSP 20252026.02 | 5.92 | — | — | — | |
| SSLVenue=ICASSP ’242026.03 | 6.1 | — | — | — | |
| MamBo-2-Hydra-N1Params (M)=316.682026.01 | 6.24 | — | — | — | |
| WaveSpecPublication=ICASSP 20252026.02 | 6.58 | — | — | — | |
| XLSR+MambaParams (M)=3192026.02 | 6.7 | — | — | — | |
| XLSR-MambaParams (M)=319.332026.01 | 6.71 | — | — | — | |
| XLSR-MambaPublication=SPL 20252026.02 | 6.71 | — | — | — | |
| XLS-R + HierConBackbone=XLS-R 300M, Head=HierCon2026.02 | 6.87 | — | — | — | |
| Wav2Vec+LogRegBackbone=Wav2Vec, Head=LogReg2026.02 | 7.2 | — | — | — | |
| XLSR+SLSParams (M)=3402026.02 | 7.45 | — | — | — | |
| XLSR-SLSParams (M)=341.492026.01 | 7.46 | — | — | — | |
| SLSPublication=ACMMM 20242026.02 | 7.46 | — | — | — | |
| Wav2Vec2+AASIST(Baseline)Params (M)=317.842026.02 | 7.58 | — | — | — | |
| OCKDPublication=ICASSP 20242026.02 | 7.68 | — | — | — | |
| XLSR+Conformer+TCMParams (M)=3192026.02 | 7.79 | — | — | — | |
| XLSR-Conformer+TCMParams (M)=319.772026.01 | 7.79 | — | — | — | |
| OCKD2026.02 | 7.86 | — | — | — | |
| XLSR+BiCrossMambaParams (M)=318.212026.02 | 7.94 | — | — | — | |
| ConformerVenue=Interspeech ’232026.03 | 8.52 | — | — | — | |
| XLS-R + SLSBackbone=XLS-R, Head=SLS2026.02 | 8.87 | — | — | — | |
| Wav2Vec2+AASISTParams (M)=317.842026.02 | 11.19 | — | — | — | |
| MoEPublication=ICASSP 20252026.02 | 12.48 | — | — | — | |
| SLIMCategory=Finetuned frontend, #Param (million)=2532024.07 | 12.5 | 89.8 | — | — | |
| SLIMPublication=NeurIPS 20242026.02 | 12.5 | — | — | — | |
| FullCategory=SLIM variants (ours), #Param (million)=112024.07 | 12.9 | 89.5 | — | — | |
| Wav2Vec+LinearBackbone=Wav2Vec, Head=Linear2026.02 | 16.17 | — | — | — | |
| SFATNet-4Data Condition=Clean (no codecs)2026.01 | 17.29 | — | 89.17 | — | |
| W2V-AASISTCategory=Finetuned frontend, #Param (million)=3172024.07 | 17.5 | 84.7 | — | — | |
| W2V-ASPCategory=Finetuned frontend, #Param (million)=3172024.07 | 18.6 | 83.6 | — | — | |
| SFAT-Net 3Data Condition=Clean (no codecs)2026.01 | 19.7 | — | 85.2 | — | |
| Dual-Granularity Orthogonal DisentanglementFine-tuning status=without fine-tuning, Variant=Full Model2026.06 | 21.58 | — | — | — | |
| WavLM-MLPPre-training (SSL-based)=true, Model Parameters=>300M, Fine-tuning status=without fine-tuning2026.06 | 21.85 | — | — | — | |
| Dual-Granularity Orthogonal Disentanglement (Cosine Only)Fine-tuning status=without fine-tuning, Variant=Cosine Only2026.06 | 22.16 | — | — | — | |
| DG-AggFine-tuning status=without fine-tuning2026.06 | 22.73 | — | — | — | |
| Wav2Vec2-AASISTPre-training (SSL-based)=true, Model Parameters=>300M, Fine-tuning status=without fine-tuning2026.06 | 23.17 | — | — | — | |
| Dual-Granularity Orthogonal Disentanglement (No Disentangle)Fine-tuning status=without fine-tuning, Variant=No Disentangle2026.06 | 23.48 | — | — | — | |
| GRL BaselineFine-tuning status=without fine-tuning2026.06 | 24.18 | — | — | — | |
| SSL-FusionCategory=Frozen frontend, #Param (million)=102024.07 | 24.2 | 76.5 | — | — | |
| W2V-LCNNCategory=Frozen frontend, #Param (million)=N/A2024.07 | 24.5 | — | — | — | |
| W2V-LLGFCategory=Frozen frontend, #Param (million)=102024.07 | 25.1 | 75.6 | — | — | |
| Enc_style+lingCategory=SLIM variants (ours), #Param (million)=102024.07 | 25.1 | 75.7 | — | — | |
| WLM-ASPCategory=Frozen frontend, #Param (million)=92024.07 | 25.4 | 75.1 | — | — | |
| DependencyCategory=SLIM variants (ours), #Param (million)=92024.07 | 25.8 | 75 | — | — | |
| RawGAT-STFine-tuning status=without fine-tuning2026.06 | 25.83 | — | — | — | |
| Dual-Granularity Orthogonal Disentanglement (Single Branch)Fine-tuning status=without fine-tuning, Variant=Single Branch2026.06 | 25.86 | — | — | — | |
| Res-TSSDNetFine-tuning status=without fine-tuning2026.06 | 26.14 | — | — | — | |
| AASISTFine-tuning status=without fine-tuning2026.06 | 27.41 | — | — | — | |
| W2V+HUBCategory=Frozen frontend, #Param (million)=92024.07 | 27.9 | 73.7 | — | — | |
| WLM+HUBCategory=Frozen frontend, #Param (million)=92024.07 | 29.2 | 72.4 | — | — | |
| Enc_styCategory=SLIM variants (ours), #Param (million)=92024.07 | 29.2 | 72.4 | — | — | |
| PS3DTCategory=Baseline, #Param (million)=N/A2024.07 | 29.7 | — | — | — | |
| HUB-ASPCategory=Frozen frontend, #Param (million)=92024.07 | 29.9 | 71.8 | — | — | |
| W2V-ASPCategory=Frozen frontend, #Param (million)=92024.07 | 30.2 | 70.5 | — | — | |
| W2V+WLMCategory=Frozen frontend, #Param (million)=92024.07 | 30.3 | 70.4 | — | — | |
| Enc_lingCategory=SLIM variants (ours), #Param (million)=92024.07 | 30.4 | 70.8 | — | — | |
| AASIST-LVenue=ICASSP ’222026.03 | 32.61 | — | — | — | |
| RawNet2Category=Baseline, #Param (million)=42024.07 | 37.8 | 60.2 | — | — | |
| AASIST-BVenue=ICASSP ’222026.03 | 39.45 | — | — | — | |
| CAFNetZero-shot=true, Trained on=MLADDC T22026.05 | 45.9 | — | 56.22 | 53.62 | |
| LCNNCategory=Baseline, #Param (million)=42024.07 | 65.6 | 37.3 | — | — | |
| AASIST-BVenue=ICASSP ’222026.03 | — | — | 65.31 | — | |
| AASIST-LVenue=ICASSP ’222026.03 | — | — | 73.72 | — | |
| AFSS2026.03 | — | — | 99.64 | — | |
| ClovaAI2026.03 | — | — | 74 | — | |
| ConformerVenue=Interspeech ’232026.03 | — | — | 97.12 | — | |
| ECAPA-TDNN2026.03 | — | — | 79 | — | |
| FlowFakeTrain=ASV192026.06 | — | — | — | 61.71 | |
| FlowFakeTrain=FoR2026.06 | — | — | — | 70.91 | |
| FlowFakeTrain=MLAAD v12026.06 | — | — | — | 62.39 | |
| H/ASP2026.03 | — | — | 85 | — | |
| HyperPotterParams (M)=317.872026.02 | — | 95.4 | — | — | |
| POI-Forensics2026.03 | — | — | 91 | — | |
| RawGAT-STTrain=ASV192026.06 | — | — | — | 50 | |
| RawGAT-STTrain=FoR2026.06 | — | — | — | 49.8 | |
| RawGAT-STTrain=MLAAD v12026.06 | — | — | — | 47.7 | |
| SCLVenue=Interspeech ’242026.03 | — | — | 99.1 | — | |
| SSL W2V2Train=ASV192026.06 | — | — | — | 79.7 | |
| SSL W2V2Train=FoR2026.06 | — | — | — | 57.8 | |
| SSL W2V2Train=MLAAD v12026.06 | — | — | — | 68 | |
| TitaNet2026.03 | — | — | 91 | — | |
| Wav2Vec2+AASISTParams (M)=317.842026.02 | — | 90.9 | — | — | |
| Wav2Vec2+AASIST (Baseline)Params (M)=317.842026.02 | — | 93.9 | — | — | |
| Whisper DFTrain=ASV192026.06 | — | — | — | 76.5 | |
| Whisper DFTrain=FoR2026.06 | — | — | — | 54.1 |