Browsing by Author "Karamatlı, Ertuğ"
Now showing items 1-4 of 4
-
Audio source separation using variational autoencoders and weak class supervision
Karamatlı, Ertuğ; Cemgil, Ali Taylan; Kırbız, Serap (Institute of Electrical and Electronics Engineers (IEEE), 2019)In this letter, we propose a source separation method that is trained by observing the mixtures and the class labels of the sources present in the mixture without any access to isolated sources. Since our method does not ... -
Joint source separation and classification using variational autoencoders
Hızlı, Çağlar; Karamatlı, Ertuğ; Kırbız, Serap (IEEE, 2020)In this paper, we propose a novel multi-task variational auto encoder (VAE) based approach for joint source separation and classification. The network uses a probabilistic encoder for each sources to map the input data to ... -
MixCycle: Unsupervised Speech Separation via Cyclic Mixture Permutation Invariant Training
Karamatlı, Ertuğ; Kırbız, Serap (IEEE, 2022)We introduce two unsupervised source separation methods, which involve self-supervised training from single-channel two-source speech mixtures. Our first method, mixture permutation invariant training (MixPIT), enables ... -
Negatif olmayan gürültü giderici değişimli oto-kodlayıcılar kullanarak tek kanaldan kaynak ayrıştırma için zayıf etiket denetimi
Karamatlı, Ertuğ; Cemgil, Ali Taylan; Kırbız, Serap (IEEE, 2019)Derin öğrenme modelleri, büyük miktarda etiketlenmiş veri bulunduğunda kaynak ayrıştırmada çok başarılı olmaktadır. Bununla birlikte, dikkatlice etiketlenmiş veri kümelerine erişim her zaman mümkün olmamaktadır. Bu bildiride, ...
