mlsanityData to 14 Oct 2026

MECT: Mixture of Experts with CNN-Transformer Network for Speaker verification

arXiv 2609.24061 ↗First posted 21 Sep 2026cs.SD cs.AI eess.AScited by 0

On the map: Mixture-of-experts → Multimodal MoE models

Abstract (arXiv)

In this paper, we propose MECT, a speaker verification model that integrates the Mixture-of-Experts (MoE) mechanism into a CNN-Transformer backbone with optimized block structure and stacking scheme. Specifically, we investigated four MoE variants that span utterance-level and frame-level granularity with dense and sparse routing strategies. The MoE mechanism proves to be effective over the baseline without MoE with only a small increase in parameters. We further scale MECT to a series of model sizes, all maintaining compact parameters and low computational complexity. In particular, MECT-B2 achieves state-of-the-art performance on VoxCeleb1 and delivers strong results on CN-Celeb, demonstrating its effectiveness across diverse datasets. In addition, we establish a streaming inference paradigm through causal retraining, which maintains strong performance at a chunk size of 100ms.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Multimodal MoE models, one of 12 directions in Mixture-of-experts. Builds on 25 of its 35 references that are in the map. First posted on arXiv 21 Sep 2026.

Builds on

25 of 35 references.

References · 25

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,509
  3. 2018VoxCeleb2: Deep Speaker Recognitioncited by 875
  4. 2020ECAPA-TDNN: Emphasized Channel Attention, Propagation and Aggregation in TDNN Based Speaker Verificationcited by 628
  5. 2015MUSAN: A Music, Speech, and Noise Corpuscited by 425
  6. 2018Attentive Statistics Pooling for Deep Speaker Embeddingcited by 114
  7. 2023CAM++: A Fast and Efficient Network for Speaker Verification Using Context-Aware Maskingcited by 102
  8. 2022Wespeaker: A Research and Production oriented Speaker Embedding Learning Toolkitcited by 97
  9. 2019BUT System Description to VoxCeleb Speaker Recognition Challenge 2019cited by 54
  10. 2019CN-CELEB: a challenging Chinese speaker recognition datasetcited by 48
  11. 2022MFA-Conformer: Multi-scale Feature Aggregation Conformer for Automatic Speaker Verificationcited by 47
  12. 2018Building and Evaluation of a Real Room Impulse Response Datasetcited by 43
13 more
  1. 2023An Enhanced Res2Net with Local and Global Feature Fusion for Speaker Verificationcited by 42
  2. 2024Reshape Dimensions Network for Speaker Recognitioncited by 28
  3. 2024VoxBlink2: A 100K+ Speaker Recognition Corpus and the Open-Set Speaker-Identification Benchmarkcited by 26
  4. 2024ECAPA2: A Hybrid Neural Network Architecture and Training Strategy for Robust Speaker Embeddingscited by 22
  5. 20243D-Speaker-Toolkit: An Open-Source Toolkit for Multimodal Speaker Verification and Diarizationcited by 16
  6. 2021Multi-View Self-Attention Based Transformer for Speaker Recognitioncited by 8
  7. 2023PCF: ECAPA-TDNN with Progressive Channel Fusion for Speaker Verificationcited by 8
  8. 2023Golden Gemini is All You Need: Finding the Sweet Spots for Speaker Verificationcited by 8
  9. 2023Exploring Binary Classification Loss For Speaker Verificationcited by 6
  10. 2023NeXt-TDNN: Modernizing Multi-Scale Temporal Convolution Backbone for Speaker Verificationcited by 6
  11. 2025Enhancing Speaker Verification with w2v-BERT 2.0 and Knowledge Distillation guided Structured Pruningcited by 3
  12. 2026ReDimNet2: Scaling Speaker Verification via Time-Pooled Dimension Reshapingcited by 3
  13. 2025Rethinking Leveraging Pre-Trained Multi-Layer Representations for Speaker Verificationcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2023Distilling Multi-Level X-vector Knowledge for Small-footprint Speaker Verification
  2. 2024Memory-Efficient Training for Deep Speaker Embedding Learning in Speaker Verification
  3. 2025Exploring Speaker Diarization with Mixture of Experts
  4. 2025Noise-Conditioned Mixture-of-Experts Framework for Robust Speaker Verification
  5. 2024MASV: Speaker Verification with Global and Local Context Mamba
  6. 2025BigMac: A Communication-Efficient Mixture-of-Experts Model Structure for Fast Training and Inference
  7. 2026FASTDIAR: Frame-level speaker encoder for Streaming Diarization
  8. 2021Class Token and Knowledge Distillation for Multi-head Self-Attention Speaker Verification Systems