Multilingually trained bottleneck features in spoken language recognition

62 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale (2021, neutral): “We compare to m-CPC [40], [41], XLSR-10 [14] and XLSR-53 [14]. respect to the read audiobook domain of CommonVoice.”
Unsupervised Cross-lingual Representation Learning for Speech Recognition (2020, neutral): “† (Fer et al., 2017) BBLall 10 1 36.”
Approaches to Multi-domain Language Recognition (2018, methodology): “The BUT-supplied bottleneck extractors [10] (BUTBNf and BUTBNo) for LRE’17 were based on LMS features with 10 stacked frames with Hamming windows applied on the trajectories prior to taking six DCT bases, to give a 144-dimensional vector that was finally normalized with mean and variance normalizati”
Speaker Invariant Feature Extraction for Zero-Resource Languages with Adversarial Learning (2018, methodology): “We therefore plan to apply the proposed multi-task learning to various kinds of neural networks such a bottleneck approach based on multilingual phonemes [23] and a Siamese network [7].”
Language learning using Speech to Image retrieval (2019, methodology): “The MBN features are created using a pre-trained DNN made available by [21].”