English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

A key challenge in machine learning is to generalize from training data to anapplication domain of interest. This work generalizes the recently-proposedmixture invariant training (MixIT) algorithm to perform unsupervised learningin the multi-channel setting. We use MixIT to train a model on far-fieldmicrophone array recordings of overlapping reverberant and noisy speech fromthe AMI Corpus. The models are trained on both supervised and unsupervisedtraining data, and are tested on real AMI recordings containing overlappingspeech. To objectively evaluate our models, we also use a syntheticmulti-channel AMI test set. Holding network architectures constant, we findthat a fine-tuned semi-supervised model yields the largest improvement toSI-SNR and to human listening ratings across synthetic and real datasets,outperforming supervised models trained on well-matched synthetic data. Ourresults demonstrate that unsupervised learning through MixIT enables modeladaptation on both single- and multi-channel real-world speech recordings.

Unsupervised Multi-channel Separation and Adaptation