An investigation of neural uncertainty estimation for target speaker extraction equipped RNN transducer

2021 ◽  
pp. 101327
Author(s):  
Jiatong Shi ◽  
Chunlei Zhang ◽  
Chao Weng ◽  
Shinji Watanabe ◽  
Meng Yu ◽  
...  
Keyword(s):  
Uncertainty Estimation ◽  
Target Speaker

Voice Conversion Across Arbitrary Speakers Based on a Single Target-Speaker Utterance

2018 ◽  
Author(s):  
Songxiang Liu ◽  
Jinghua Zhong ◽  
Lifa Sun ◽  
Xixin Wu ◽  
Xunying Liu ◽  
...  
Keyword(s):  
Voice Conversion ◽  
Single Target ◽  
Target Speaker

Development and Certification of Formononetin Reference Material for Quality Control of Functional Foods and Botanical Supplements

2019 ◽  
Vol 15 (5) ◽  
pp. 553-559
Author(s):  
Ningbo Gong ◽  
Baoxi Zhang ◽  
Kun Hu ◽  
Zhaolin Gao ◽  
Guanhua Du ◽  
...  
Keyword(s):  
Quality Control ◽  
Reference Material ◽  
Functional Foods ◽  
Certified Reference Materials ◽  
Coulometric Titration ◽  
Pharmaceutical Formulations ◽  
Uncertainty Estimation ◽  
Expanded Uncertainty ◽  
Scanning Calorimetry ◽  
Botanical Supplements

Background: Formononetin is a common soy isoflavonoid that can be found abundantly in many natural plants. Previous studies have shown that formononetin possesses a variety of activities which can be applied for various medicinal purposes. Certified Reference Materials (CRMs) play a fundamental role in the food, traditional medicine and dietary supplement fields, and can be used for method validation, uncertainty estimation, as well as quality control. Methods: The purity of formononetin was determined by Differential Scanning Calorimetry (DSC), Coulometric Titration (CT) and Mass Balance (MB) methods. Results: This paper reports the sample preparation methodology, homogeneity and stability studies, value assignment, and uncertainty estimation of a new certified reference material of formononetin. DSC, CT and MB methods proved to be sufficiently reliable and accurate for the certification purpose. The purity of the formononetin CRM was therefore found to be 99.40% ± 0.24 % (k = 2) based on the combined value assignments and the expanded uncertainty. Conclusion: This CRM will be a reliable standard for the validation of the analytical methods and for quality assurance/quality control of formononetin and formononetin-related traditional herbs, food products, dietary supplements and pharmaceutical formulations.

scholarly journals Calibration and uncertainty estimation of non-contact coordinate measurement systems based on Kriging models

2019 ◽  
Vol 57 ◽  
pp. 16-29 ◽  
Author(s):  
Lan Fei ◽  
Jean-Yves Dantan ◽  
Cyrille Baudouin ◽  
Shichang Du
Keyword(s):  
Uncertainty Estimation ◽  
Coordinate Measurement ◽  
Measurement Systems

IMPROVING EIGENSPACE-BASED FUZZY LOGIC SYSTEM USING A LINEAR INTERPOLATION SCHEME FOR SPEECH PATTERN RECOGNITION

2013 ◽  
Vol 37 (3) ◽  
pp. 611-620
Author(s):  
Ing-Jr Ding ◽  
Chih-Ta Yen
Keyword(s):  
Pattern Recognition ◽  
Fuzzy Logic ◽  
Recognition Performance ◽  
Linear Interpolation ◽  
Fuzzy Logic System ◽  
Interpolation Scheme ◽  
Speech Pattern ◽  
Eigen Value ◽  
Target Speaker ◽  
Logic System

The Eigen-FLS approach using an eigenspace-based scheme for fast fuzzy logic system (FLS) establishments has been attempted successfully in speech pattern recognition. However, speech pattern recognition by Eigen-FLS will still encounter a dissatisfactory recognition performance when the collected data for eigen value calculations of the FLS eigenspace is scarce. To tackle this issue, this paper proposes two improved-versioned Eigen-FLS methods, incremental MLED Eigen-FLS and EigenMLLR-like Eigen-FLS, both of which use a linear interpolation scheme for properly adjusting the target speaker’s Eigen-FLS model derived from an FLS eigenspace. Developed incremental MLED Eigen-FLS and EigenMLLR-like Eigen-FLS are superior to conventional Eigen-FLS especially in the situation of insufficient data from the target speaker.

scholarly journals Quantifying the uncertainty of deep learning-based computer-aided diagnosis for patient safety

2019 ◽  
Vol 5 (1) ◽  
pp. 223-226
Author(s):  
Max-Heinrich Laves ◽  
Sontje Ihler ◽  
Tobias Ortmaier ◽  
Lüder A. Kahrs
Keyword(s):  
Monte Carlo ◽  
Patient Safety ◽  
Deep Learning ◽  
Epistemic Uncertainty ◽  
Computer Aided Diagnosis ◽  
Uncertainty Estimation ◽  
Test Time ◽  
Prediction Uncertainty ◽  
Computer Aided ◽  
Aided Diagnosis

AbstractIn this work, we discuss epistemic uncertainty estimation obtained by Bayesian inference in diagnostic classifiers and show that the prediction uncertainty highly correlates with goodness of prediction. We train the ResNet-18 image classifier on a dataset of 84,484 optical coherence tomography scans showing four different retinal conditions. Dropout is added before every building block of ResNet, creating an approximation to a Bayesian classifier. Monte Carlo sampling is applied with dropout at test time for uncertainty estimation. In Monte Carlo experiments, multiple forward passes are performed to get a distribution of the class labels. The variance and the entropy of the distribution is used as metrics for uncertainty. Our results show strong correlation with ρ = 0.99 between prediction uncertainty and prediction error. Mean uncertainty of incorrectly diagnosed cases was significantly higher than mean uncertainty of correctly diagnosed cases. Modeling of the prediction uncertainty in computer-aided diagnosis with deep learning yields more reliable results and is therefore expected to increase patient safety. This will help to transfer such systems into clinical routine and to increase the acceptance of machine learning in diagnosis from the standpoint of physicians and patients.

PO-1716: Histology correlation of in-vivo [68Ga]PSMAPET/MRI: A method description and uncertainty estimation

2020 ◽  
Vol 152 ◽  
pp. S948
Author(s):  
K. Sandgren ◽  
J. Jonsson ◽  
A. Keeratijarut Lindberg ◽  
T. Näsmark ◽  
S. Said ◽  
...  
Keyword(s):  
Uncertainty Estimation
Sign in / Sign up

Export Citation Format

Share Document