Instabilities in Convnets for Raw Audio
What makes waveform-based deep learning so hard? Despite numerous attempts at training convolutional neural networks (convnets) for filterbank design, they often fail to outperform hand-crafted baselines. These baselines are linear time-invariant systems: as such, they can be approximated by convnets with wide receptive fields. Yet, in practice, gradient-based optimization leads to suboptimal approximations. In our article, we approach this phenomenon from the perspective of initialization. We present a theory of large deviations for the energy response of FIR filterbanks with random Gaussian weights. We find that deviations worsen for large filters and locally periodic input signals, which are both typical for audio signal processing applications. Numerical simulations align with our theory and suggest that the condition number of a convolutional layer follows a logarithmic scaling law between the number and length of the filters, which is reminiscent of discrete wavelet bases.
Code (1)
Tasks
Audio Signal ProcessingMethods 이 논문이 사용한 방법론
Similar Papers 제목 키워드 기반
UniRepLKNet: A Universal Perception Large-Kernel ConvNet for Audio Video Point Cloud Time-Series and Image Recognition
Large-kernel convolutional neural networks (ConvNets) have recently received extensive research attention but two unresolved and critical issues demand further investigation. 1) The architectures of existing large-ke…
Time SeriesTime Series ForecastingUniRepLKNet: A Universal Perception Large-Kernel ConvNet for Audio, Video, Point Cloud, Time-Series and Image Recognition
Large-kernel convolutional neural networks (ConvNets) have recently received extensive research attention, but two unresolved and critical issues demand further investigation. 1) The architectures of existing large-kerne…
Image ClassificationObject DetectionSemantic SegmentationTime Series+1Spatio-Temporal Graph Convolutional Neural Networks for Physics-Aware Grid Learning Algorithms
This paper proposes a model-free Volt-VAR control (VVC) algorithm via the spatio-temporal graph ConvNet-based deep reinforcement learning (STGCN-DRL) framework, whose goal is to control smart inverters in an unbalanced d…
Deep Reinforcement Learningreinforcement-learningReinforcement LearningReinforcement Learning (RL)Scaling Up Your Kernels: Large Kernel Design in ConvNets towards Universal Representations
This paper proposes the paradigm of large convolutional kernels in designing modern Convolutional Neural Networks (ConvNets). We establish that employing a few large kernels, instead of stacking multiple smaller ones, ca…
Time Series ForecastingVideo RecognitionHold Me Tight: Stable Encoder-Decoder Design for Speech Enhancement
Convolutional layers with 1-D filters are often used as frontend to encode audio signals. Unlike fixed time-frequency representations, they can adapt to the local characteristics of input data. However, 1-D filters on ra…
DecoderSpeech Enhancement