M
Miho Ohsaki
Researcher at Doshisha University
Publications - 6
Citations - 18
Miho Ohsaki is an academic researcher from Doshisha University. The author has contributed to research in topics: Bayes' theorem & Posterior probability. The author has an hindex of 3, co-authored 6 publications receiving 15 citations.
Papers
More filters
Journal ArticleDOI
A Practical Method Based on Bayes Boundary-Ness for Optimal Classifier Parameter Status Selection
TL;DR: A novel practical method for finding the optimal classifier parameter status corresponding to the Bayes error through the evaluation of estimated class boundaries from the perspective of Bayes boundary-ness with an entropy-based uncertainty measure is proposed.
Proceedings ArticleDOI
Optimal classifier model status selection using bayes boundary uncertainty
David Ha,Emilie Delattre,Yuya Tomotoshi,Masahiro Senda,Hideyuki Watanabe,Shigeru Katagiri,Miho Ohsaki +6 more
TL;DR: This work proposes a method to select the optimal parameter status for any classifier model by evaluating the ideality of the classifier’s classification boundary instead of estimating the error probability, using the fact that the Bayes boundary solely consists of uncertain samples.
Book ChapterDOI
Optimal Classifier Parameter Status Selection Based on Bayes Boundary-ness for Multi-ProtoType and Multi-Layer Perceptron Classifiers
Yuya Tomotoshi,David Ha,Emilie Delattre,Hideyuki Watanabe,Xugang Lu,Shigeru Katagiri,Miho Ohsaki +6 more
TL;DR: The proposed method to select an optimal classifier parameter status (value) using the new criterion that is referred to as uncertainty measure and directly evaluates the Bayes boundary-ness of estimated boundaries can be applied to two types of classifiers whose class boundaries are basically nonlinear.
Proceedings ArticleDOI
Maximum Bayes Boundary-Ness Training For Pattern Classification
TL;DR: This work proposes a new one-stage training method that directly optimizes a given classifier parameter set by maximizing its Bayes boundary-ness or increasing its accuracy during Bayes error estimation.
Proceedings ArticleDOI
Minimum Classification Error Training with Speech Synthesis-Based Regularization for Speech Recognition
TL;DR: The quality of the synthesized speech using LSP parameters derived from the trained prototypes is found to be clearly supported by the speech synthesis ability preserved in the training, and thequality of the Bayes error estimation is clearly supported.