Enhance Curvature Information by Structured Stochastic Quasi-Newton Methods

no code implementations • CVPR 2021 • Ming-Han Yang, Dong Xu, Hongyu Chen, Zaiwen Wen, Mengyun Chen

In this paper, we consider stochastic second-order methods for minimizing a finite summation of nonconvex functions.

Second-order methods

Sketchy Empirical Natural Gradient Methods for Deep Learning

1 code implementation • 10 Jun 2020 • Ming-Han Yang, Dong Xu, Zaiwen Wen, Mengyun Chen, Pengxiang Xu

Experiments on the distributed large-batch training show that the scaling efficiency is quite reasonable.

A Stochastic Extra-Step Quasi-Newton Method for Nonsmooth Nonconvex Optimization

no code implementations • 21 Oct 2019 • Ming-Han Yang, Andre Milzarek, Zaiwen Wen, Tong Zhang

In this paper, a novel stochastic extra-step quasi-Newton method is developed to solve a class of nonsmooth nonconvex composite optimization problems.

融合多任務學習類神經網路聲學模型訓練於會議語音辨識之研究 (Leveraging Multi-Task Learning with Neural Network Based Acoustic Modeling for Improved Meeting Speech Recognition) [In Chinese]

no code implementations • ROCLINGIJCLCLP 2016 • Ming-Han Yang, Yao-Chi Hsu, Hsiao-Tsung Hung, Ying-Wen Chen, Kuan-Yu Chen, Berlin Chen

Multi-Task Learning speech-recognition +1

評估尺度相關最佳化方法於華語錯誤發音檢測之研究 (Evaluation Metric-related Optimization Methods for Mandarin Mispronunciation Detection) [In Chinese]

no code implementations • ROCLINGIJCLCLP 2016 • Yao-Chi Hsu, Ming-Han Yang, Hsiao-Tsung Hung, Yi-Ju Lin, Kuan-Yu Chen, Berlin Chen

評估尺度相關最佳化方法於華語錯誤發音檢測之研究(Evaluation Metric-related Optimization Methods for Mandarin Mispronunciation Detection) [In Chinese]

no code implementations • ROCLINGIJCLCLP 2016 • Yao-Chi Hsu, Ming-Han Yang, Hsiao-Tsung Hung, Yi-Ju Lin, Berlin Chen

融合多任務學習類神經網路聲學模型訓練於會議語音辨識之研究(Leveraging Multi-task Learning with Neural Network Based Acoustic Modeling for Improved Meeting Speech Recognition) [In Chinese]

no code implementations • ROCLINGIJCLCLP 2016 • Ming-Han Yang, Yao-Chi Hsu, Hsiao-Tsung Hung, Ying-Wen Chen, Berlin Chen, Kuan-Yu Chen

Multi-Task Learning speech-recognition +1

融合多種深層類神經網路聲學模型與分類技術於華語錯誤發音檢測之研究(Exploring Combinations of Various Deep Neural Network based Acoustic Models and Classification Techniques for Mandarin Mispro-nunciation Detection)[In Chinese]

no code implementations • ROCLINGIJCLCLP 2015 • Yao-Chi Hsu, Ming-Han Yang, Hsiao-Tsung Hung, Yuwen Hsiung, Yao-Ting Hung, Berlin Chen