# Improving Sound Event Classification by Increasing Shift Invariance in Convolutional Neural Networks

arXiv: https://arxiv.org/abs/2107.00623 · HTML: https://mlsanity.com/paper/2107.00623/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 33 references that are in the map.
- First posted on arXiv 1 Jul 2021.

## Builds on (most cited first, 22 of 22)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https://mlsanity.com/paper/1502.03167/) (2015)
- [mixup: Beyond Empirical Risk Minimization](https://mlsanity.com/paper/1710.09412/) (2017)
- [CNN Architectures for Large-Scale Audio Classification](https://mlsanity.com/paper/1609.09430/) (2016)
- [PANNs: Large-Scale Pretrained Audio Neural Networks for Audio Pattern Recognition](https://mlsanity.com/paper/1912.10211/) (2019)
- [Making Convolutional Networks Shift-Invariant Again](https://mlsanity.com/paper/1904.11486/) (2019)
- [FSD50K: An Open Dataset of Human-Labeled Sound Events](https://mlsanity.com/paper/2010.00475/) (2020)
- [Harmonic Networks: Deep Translation and Rotation Equivariance](https://mlsanity.com/paper/1612.04642/) (2016)
- [Why do deep convolutional networks generalize so poorly to small image transformations?](https://mlsanity.com/paper/1805.12177/) (2018)
- [Exploiting Cyclic Symmetry in Convolutional Neural Networks](https://mlsanity.com/paper/1602.02660/) (2016)
- [Exploring the Landscape of Spatial Robustness](https://mlsanity.com/paper/1712.02779/) (2017)
- [LEAF: A Learnable Frontend for Audio Classification](https://mlsanity.com/paper/2101.08596/) (2021)
- [Truly shift-invariant convolutional neural networks](https://mlsanity.com/paper/2011.14214/) (2020)
- [Evaluation of CNN-based Automatic Music Tagging Models](https://mlsanity.com/paper/2006.00751/) (2020)
- [Audio Transformers](https://mlsanity.com/paper/2105.00335/) (2021)
- [Unsupervised Contrastive Learning of Sound Event Representations](https://mlsanity.com/paper/2011.07616/) (2020)
- [An Effective Anti-Aliasing Approach for Residual Networks](https://mlsanity.com/paper/2011.10675/) (2020)
- [Learning Sound Event Classifiers from Web Audio with Noisy Labels](https://mlsanity.com/paper/1901.01189/) (2019)
- [Addressing Missing Labels in Large-Scale Sound Event Recognition Using a Teacher-Student Framework With Loss Masking](https://mlsanity.com/paper/2005.00878/) (2020)
- [End-to-End Auditory Object Recognition via Inception Nucleus](https://mlsanity.com/paper/2005.12195/) (2020)
- [Model-agnostic Approaches to Handling Noisy Labels When Training Sound Event Classifiers](https://mlsanity.com/paper/1910.12004/) (2019)

## Built on by (most cited first, 5 of 5)

- [HEAR: Holistic Evaluation of Audio Representations](https://mlsanity.com/paper/2203.03022/) (2022)
- [Learning strides in convolutional neural networks](https://mlsanity.com/paper/2202.01653/) (2022)
- [CMKD: CNN/Transformer-Based Cross-Model Knowledge Distillation for Audio Classification](https://mlsanity.com/paper/2203.06760/) (2022)
- [Heterogeneous sound classification with the Broad Sound Taxonomy and Dataset](https://mlsanity.com/paper/2410.00980/) (2024)
- [AaSP: Aliasing-aware Self-Supervised Pre-Training for Audio Spectrogram Transformers](https://mlsanity.com/paper/2512.03637/) (2025)

## Abstract (arXiv)

Recent studies have put into question the commonly assumed shift invariance property of convolutional networks, showing that small shifts in the input can affect the output predictions substantially. In this paper, we analyze the benefits of addressing lack of shift invariance in CNN-based sound event classification. Specifically, we evaluate two pooling methods to improve shift invariance in CNNs, based on low-pass filtering and adaptive sampling of incoming feature maps. These methods are implemented via small architectural modifications inserted into the pooling layers of CNNs. We evaluate the effect of these architectural changes on the FSD50K dataset using models of different capacity and in presence of strong regularization. We show that these modifications consistently improve sound event classification in all cases considered. We also demonstrate empirically that the proposed pooling methods increase shift invariance in the network, making it more robust against time/frequency shifts in input spectrograms. This is achieved by adding a negligible amount of trainable parameters, which makes these methods an appealing alternative to conventional pooling layers. The outcome is a new state-of-the-art mAP of 0.541 on the FSD50K classification benchmark.