Neural network alternatives toconvolutive audio models for source separation

Venkataramani, Shrikant; Subakan, Cem; Smaragdis, Paris

doi:10.1109/mlsp.2017.8168108

Cited by 12 publications

(9 citation statements)

References 11 publications

(16 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recent work has explored the connection between CDL and auto-encoders in deep learning [5,6]. Framing dictionary learning as training of a neural network opens the possibility of exploiting the GPU-based infrastructure that has been developed for training neural networks, to reduce computational requirements and decrease runtime.…”

Section: Introductionmentioning

confidence: 99%

Scalable Convolutional Dictionary Learning With Constrained Recurrent Sparse Auto-Encoders

Tolooshams

Dey²,

2018

2018 IEEE 28th International Workshop on Machine Learning for Signal Processing (MLSP)

View full text Add to dashboard Cite

Given a convolutional dictionary underlying a set of observed signals, can a carefully designed auto-encoder recover the dictionary in the presence of noise? We introduce an auto-encoder architecture, termed constrained recurrent sparse auto-encoder (CRsAE), that answers this question in the affirmative. Given an input signal and an approximate dictionary, the encoder finds a sparse approximation using FISTA. The decoder reconstructs the signal by applying the dictionary to the output of the encoder. The encoder and decoder in CRsAE parallel the sparse-coding and dictionary update steps in optimization-based alternating-minimization schemes for dictionary learning. As such, the parameters of the encoder and decoder are not independent, a constraint which we enforce for the first time. We derive the backpropagation algorithm for CRsAE. CRsAE is a framework for blind source separation that, only knowing the number of sources (dictionary elements), and assuming sparsely-many can overlap, is able to separate them. We demonstrate its utility in the context of spike sorting, a source separation problem in computational neuroscience. We demonstrate the ability of CRsAE to recover the underlying dictionary and characterize its sensitivity as a function of SNR.

show abstract

Section: Introductionmentioning

confidence: 99%

Scalable Convolutional Dictionary Learning With Constrained Recurrent Sparse Auto-Encoders

Tolooshams

Dey²,

2018

2018 IEEE 28th International Workshop on Machine Learning for Signal Processing (MLSP)

View full text Add to dashboard Cite

show abstract

“…This allows the network to learn a non-negative H and a non-negative reconstruction S of the input S. Unlike NMF, the decoder weights need not be strictly non-negative. But, under suitable sparsity constraints on the activations H, they can be shown to be non-negative like NMF bases [19,20].…”

Section: Non-negative Autoencodersmentioning

confidence: 99%

“…We can now take advantage of the modeling flexibility of neural networks and develop complex encoder and decoder architectures that adhere to the above format. In particular, multilayer [19] and convolutional extensions [20] have shown significant performance improvement compared to a single dense-layer encoder and decoder given in Eq. 1.…”

Section: Non-negative Autoencodersmentioning

confidence: 99%

See 1 more Smart Citation

End-To-End Non-Negative Autoencoders for Sound Source Separation

Venkataramani

Tzinis

Smaragdis

2020

ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

Self Cite

View full text Add to dashboard Cite

Discriminative models for source separation have recently been shown to produce impressive results. However, when operating on sources outside of the training set, these models can not perform as well and are cumbersome to update. Classical methods like Nonnegative Matrix Factorization (NMF) provide modular approaches to source separation that can be easily updated to adapt to new mixture scenarios. In this paper, we generalize NMF to develop end-to-end non-negative auto-encoders and demonstrate how they can be used for source separation. Our experiments indicate that these models deliver comparable separation performance to discriminative approaches, while retaining the modularity of NMF and the modeling flexibility of neural networks.

show abstract

“…Recently, this class of factorization techniques has started to fall under the deep learning umbrella. These methods interpret non-negative autoencoders as an extension of non-negative factorization methods [5,6]. These hybrid models benefit from the advances in deep learning while maintaining the interpretability that many deep models forgo.…”

Section: Introductionmentioning

confidence: 99%

Deep Tensor Factorization for Spatially-Aware Scene Decomposition

Casebeer

Colomb

Smaragdis

2019

2019 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA)

Self Cite

View full text Add to dashboard Cite

We propose a completely unsupervised method to understand audio scenes observed with random microphone arrangements by decomposing the scene into its constituent sources and their relative presence in each microphone. To this end, we formulate a neural network architecture that can be interpreted as a nonnegative tensor factorization of a multi-channel audio recording. By clustering on the learned network parameters corresponding to channel content, we can learn sources' individual spectral dictionaries and their activation patterns over time. Our method allows us to leverage deep learning advances like end-to-end training, while also allowing stochastic minibatch training so that we can feasibly decompose realistic audio scenes that are intractable to decompose using standard methods. This neural network architecture is easily extensible to other kinds of tensor factorizations.

show abstract

Neural network alternatives toconvolutive audio models for source separation

Cited by 12 publications

References 11 publications

Scalable Convolutional Dictionary Learning With Constrained Recurrent Sparse Auto-Encoders

Scalable Convolutional Dictionary Learning With Constrained Recurrent Sparse Auto-Encoders

End-To-End Non-Negative Autoencoders for Sound Source Separation

Deep Tensor Factorization for Spatially-Aware Scene Decomposition

Contact Info

Product

Resources

About