TY - GEN
T1 - Adaptive denoising autoencoders
T2 - 12th International Conference on Latent Variable Analysis and Signal Separation, LVA/ICA 2015
AU - Kim, Minje
AU - Smaragdis, Paris
N1 - Publisher Copyright:
© Springer International Publishing Switzerland 2015.
PY - 2015
Y1 - 2015
N2 - This work aims at a test-time fine-tune scheme to further improve the performance of an already-trained Denoising AutoEncoder (DAE) in the context of semi-supervised audio source separation. Although the state-of-the-art deep learning-based DAEs show sensible denoising performance when the nature of artifacts is known in advance, the scalability of an already-trained network to an unseen signal with an unknown characteristic of deformation is not well studied. To handle this problem, we propose an adaptive fine-tuning scheme where we define a test-time target variables so that a DAE can learn from the newly available sources and the mixing environments in the test mixtures. In the proposed network topology, we stack an AutoEncoder (AE) trained from clean source spectra of interest on top of a DAE trained from a variety of available mixture spectra. Hence, the bottom DAE outputs are used as the input to the top AE, which is to check the purity of the once denoised DAE output. Then, the top AE error is used to fine-tune the bottom DAE during the test phase. Experimental results on audio source separation tasks demonstrate that the proposed fine-tuning technique can further improve the sound quality of a DAE during the test procedure.
AB - This work aims at a test-time fine-tune scheme to further improve the performance of an already-trained Denoising AutoEncoder (DAE) in the context of semi-supervised audio source separation. Although the state-of-the-art deep learning-based DAEs show sensible denoising performance when the nature of artifacts is known in advance, the scalability of an already-trained network to an unseen signal with an unknown characteristic of deformation is not well studied. To handle this problem, we propose an adaptive fine-tuning scheme where we define a test-time target variables so that a DAE can learn from the newly available sources and the mixing environments in the test mixtures. In the proposed network topology, we stack an AutoEncoder (AE) trained from clean source spectra of interest on top of a DAE trained from a variety of available mixture spectra. Hence, the bottom DAE outputs are used as the input to the top AE, which is to check the purity of the once denoised DAE output. Then, the top AE error is used to fine-tune the bottom DAE during the test phase. Experimental results on audio source separation tasks demonstrate that the proposed fine-tuning technique can further improve the sound quality of a DAE during the test procedure.
KW - Autoencoders
KW - Deep learning
KW - Deep neural networks
KW - Semi-supervised separation
KW - Speech enhancement
UR - http://www.scopus.com/inward/record.url?scp=84944679475&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84944679475&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-22482-4_12
DO - 10.1007/978-3-319-22482-4_12
M3 - Conference contribution
AN - SCOPUS:84944679475
SN - 9783319224817
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 100
EP - 107
BT - Latent Variable Analysis and Signal Separation - 12th International Conference, LVA/ICA 2015, Proceedings
A2 - Koldovský, Zbynĕk
A2 - Vincent, Emmanuel
A2 - Yeredor, Arie
A2 - Tichavský, Petr
PB - Springer
Y2 - 25 August 2015 through 28 August 2015
ER -