mlsanityData to 7 Oct 2026

Domain Generalization with Relaxed Instance Frequency-wise Normalization for Multi-device Acoustic Scene Classification

arXiv 2206.12513 ↗First posted 24 Jun 2022cs.SD cs.LG eess.AScited by 15

Not on the map. Why?

Abstract (arXiv)

While using two-dimensional convolutional neural networks (2D-CNNs) in image processing, it is possible to manipulate domain information using channel statistics, and instance normalization has been a promising way to get domain-invariant features. However, unlike image processing, we analyze that domain-relevant information in an audio feature is dominant in frequency statistics rather than channel statistics. Motivated by our analysis, we introduce Relaxed Instance Frequency-wise Normalization (RFN): a plug-and-play, explicit normalization module along the frequency axis which can eliminate instance-specific domain discrepancy in an audio feature while relaxing undesirable loss of useful discriminative information. Empirically, simply adding RFN to networks shows clear margins compared to previous domain generalization approaches on acoustic scene classification and yields improved robustness for multiple audio devices. Especially, the proposed RFN won the DCASE2021 challenge TASK1A, low-complexity acoustic scene classification with multiple devices, with a clear margin, and RFN is an extended work of our technical report.

Key facts. Cited by 15 papers in the map (data through 7 Oct 2026). Builds on 32 of its 39 references that are in the map. First posted on arXiv 24 Jun 2022.

Builds on

32 of 39 references.

References · 32

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
  3. 2016Layer Normalizationcited by 5,579
  4. 2017mixup: Beyond Empirical Risk Minimizationcited by 5,155
  5. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 4,993
  6. 2017Arbitrary Style Transfer in Real-time with Adaptive Instance Normalizationcited by 1,998
  7. 2018Group Normalizationcited by 1,514
  8. 2017Accurate, Large Minibatch SGD: Training ImageNet in 1 Hourcited by 1,479
  9. 2019SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognitioncited by 1,355
  10. 2016Instance Normalization: The Missing Ingredient for Fast Stylizationcited by 1,187
  11. 2017Deeper, Broader and Artier Domain Generalizationcited by 983
  12. 2018Speech Commands: A Dataset for Limited-Vocabulary Speech Recognitioncited by 741
20 more
  1. 2016CNN Architectures for Large-Scale Audio Classificationcited by 701
  2. 2021Domain Generalization with MixStylecited by 563
  3. 2018Two at Once: Enhancing Learning and Generalization Capacities via IBN-Netcited by 361
  4. 2018Additive Margin Softmax for Face Verificationcited by 275
  5. 2018Generalizing Across Domains via Cross-Gradient Trainingcited by 268
  6. 2017Improved Texture Networks: Maximizing Quality and Diversity in Feed-forward Stylization and Texture Synthesiscited by 215
  7. 2020Efficient Domain Generalization via Common-Specific Low-Rank Decompositioncited by 134
  8. 2020In defence of metric learning for speaker recognitioncited by 123
  9. 2018Batch-Instance Normalization for Adaptively Style-Invariant Neural Networkscited by 77
  10. 2020Meta Batch-Instance Normalization for Generalizable Person Re-Identificationcited by 77
  11. 2021Revisiting Locally Supervised Learning: an Alternative to End-to-end Trainingcited by 71
  12. 2020TaskNorm: Rethinking Batch Normalization for Meta-Learningcited by 50
  13. 2020Acoustic scene classification in DCASE 2020 Challenge: generalization across devices and low complexity solutionscited by 48
  14. 2020Rethinking CNN Models for Audio Classificationcited by 48
  15. 2019CN-CELEB: a challenging Chinese speaker recognition datasetcited by 47
  16. 2021Broadcasted Residual Learning for Efficient Keyword Spottingcited by 47
  17. 2016Trainable Frontend For Robust and Far-Field Keyword Spottingcited by 32
  18. 2020CN-Celeb: multi-genre speaker recognitioncited by 25
  19. 2018SubSpectralNet - Using Sub-Spectrogram based Convolutional Neural Networks for Acoustic Scene Classificationcited by 10
  20. 2020Domain-Invariant Speaker Vector Projection by Model-Agnostic Meta-Learningcited by 3

Built on by

Cited by year

Citing papers per year: 2022: 3; 2023: 5; 2024: 3; 2025: 4.

15 papers.

Sort:

  1. 2023TTN: A Domain-Shift Aware Batch Normalization in Test-Time Adaptationcited by 101
  2. 2022QTI Submission to DCASE 2021: residual normalization for device-imbalanced acoustic scene classification with efficient designcited by 10
  3. 2023Device-Robust Acoustic Scene Classification via Impulse Response Augmentationcited by 10
  4. 2025TACOS: Temporally-aligned Audio CaptiOnS for Language-Audio Pretrainingcited by 9
  5. 2022Dummy Prototypical Networks for Few-Shot Open-Set Keyword Spottingcited by 7
  6. 2022Improving Natural-Language-based Audio Retrieval with Transfer Learning and Audio & Text Augmentationscited by 2
  7. 2023Domain Information Control at Inference Time for Acoustic Scene Classificationcited by 1
  8. 2024Microphone Conversion: Mitigating Device Variability in Sound Event Classificationcited by 1
  9. 2024Unified Microphone Conversion: Many-to-Many Device Mapping via Feature-wise Linear Modulationcited by 1
  10. 2023On Frequency-Wise Normalizations for Better Recording Device Generalization in Audio Spectrogram Transformerscited by 0
  11. 2023Improving Small Footprint Few-shot Keyword Spotting with Supervision on Auxiliary Datacited by 0
  12. 2024Data Efficient Acoustic Scene Classification using Teacher-Informed Confusing Class Instructioncited by 0
3 more
  1. 2025PatchDSU: Uncertainty Modeling for Out of Distribution Generalization in Keyword Spottingcited by 0
  2. 2025Adaptive Knowledge Distillation using a Device-Aware Teacher for Low-Complexity Acoustic Scene Classificationcited by 0
  3. 2025Lightweight and Generalizable Acoustic Scene Representations via Contrastive Fine-Tuning and Distillationcited by 0