Alert button
Picture for Yaoyu Zhang

Yaoyu Zhang

Alert button

Structure and Gradient Dynamics Near Global Minima of Two-layer Neural Networks

Sep 01, 2023
Leyang Zhang, Yaoyu Zhang, Tao Luo

Viaarxiv icon

Optimistic Estimate Uncovers the Potential of Nonlinear Models

Jul 18, 2023
Yaoyu Zhang, Zhongwang Zhang, Leyang Zhang, Zhiwei Bai, Tao Luo, Zhi-Qin John Xu

Figure 1 for Optimistic Estimate Uncovers the Potential of Nonlinear Models
Figure 2 for Optimistic Estimate Uncovers the Potential of Nonlinear Models
Figure 3 for Optimistic Estimate Uncovers the Potential of Nonlinear Models
Figure 4 for Optimistic Estimate Uncovers the Potential of Nonlinear Models
Viaarxiv icon

Linear Stability Hypothesis and Rank Stratification for Nonlinear Models

Nov 21, 2022
Yaoyu Zhang, Zhongwang Zhang, Leyang Zhang, Zhiwei Bai, Tao Luo, Zhi-Qin John Xu

Figure 1 for Linear Stability Hypothesis and Rank Stratification for Nonlinear Models
Figure 2 for Linear Stability Hypothesis and Rank Stratification for Nonlinear Models
Figure 3 for Linear Stability Hypothesis and Rank Stratification for Nonlinear Models
Figure 4 for Linear Stability Hypothesis and Rank Stratification for Nonlinear Models
Viaarxiv icon

Embedding Principle in Depth for the Loss Landscape Analysis of Deep Neural Networks

May 26, 2022
Zhiwei Bai, Tao Luo, Zhi-Qin John Xu, Yaoyu Zhang

Figure 1 for Embedding Principle in Depth for the Loss Landscape Analysis of Deep Neural Networks
Figure 2 for Embedding Principle in Depth for the Loss Landscape Analysis of Deep Neural Networks
Figure 3 for Embedding Principle in Depth for the Loss Landscape Analysis of Deep Neural Networks
Figure 4 for Embedding Principle in Depth for the Loss Landscape Analysis of Deep Neural Networks
Viaarxiv icon

Empirical Phase Diagram for Three-layer Neural Networks with Infinite Width

May 24, 2022
Hanxu Zhou, Qixuan Zhou, Zhenyuan Jin, Tao Luo, Yaoyu Zhang, Zhi-Qin John Xu

Figure 1 for Empirical Phase Diagram for Three-layer Neural Networks with Infinite Width
Figure 2 for Empirical Phase Diagram for Three-layer Neural Networks with Infinite Width
Figure 3 for Empirical Phase Diagram for Three-layer Neural Networks with Infinite Width
Figure 4 for Empirical Phase Diagram for Three-layer Neural Networks with Infinite Width
Viaarxiv icon

Limitation of characterizing implicit regularization by data-independent functions

Jan 28, 2022
Leyang Zhang, Zhi-Qin John Xu, Tao Luo, Yaoyu Zhang

Figure 1 for Limitation of characterizing implicit regularization by data-independent functions
Figure 2 for Limitation of characterizing implicit regularization by data-independent functions
Viaarxiv icon

Overview frequency principle/spectral bias in deep learning

Jan 19, 2022
Zhi-Qin John Xu, Yaoyu Zhang, Tao Luo

Figure 1 for Overview frequency principle/spectral bias in deep learning
Figure 2 for Overview frequency principle/spectral bias in deep learning
Figure 3 for Overview frequency principle/spectral bias in deep learning
Figure 4 for Overview frequency principle/spectral bias in deep learning
Viaarxiv icon

A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics

Jan 09, 2022
Tianhan Zhang, Yuxiao Yi, Yifan Xu, Zhi X. Chen, Yaoyu Zhang, Weinan E, Zhi-Qin John Xu

Figure 1 for A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics
Figure 2 for A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics
Figure 3 for A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics
Figure 4 for A multi-scale sampling method for accurate and robust deep neural network to predict combustion chemical kinetics
Viaarxiv icon

A deep learning-based model reduction (DeePMR) method for simplifying chemical kinetics

Jan 07, 2022
Zhiwei Wang, Yaoyu Zhang, Yiguang Ju, Weinan E, Zhi-Qin John Xu, Tianhan Zhang

Figure 1 for A deep learning-based model reduction (DeePMR) method for simplifying chemical kinetics
Figure 2 for A deep learning-based model reduction (DeePMR) method for simplifying chemical kinetics
Figure 3 for A deep learning-based model reduction (DeePMR) method for simplifying chemical kinetics
Figure 4 for A deep learning-based model reduction (DeePMR) method for simplifying chemical kinetics
Viaarxiv icon

Embedding Principle: a hierarchical structure of loss landscape of deep neural networks

Nov 30, 2021
Yaoyu Zhang, Yuqing Li, Zhongwang Zhang, Tao Luo, Zhi-Qin John Xu

Figure 1 for Embedding Principle: a hierarchical structure of loss landscape of deep neural networks
Figure 2 for Embedding Principle: a hierarchical structure of loss landscape of deep neural networks
Figure 3 for Embedding Principle: a hierarchical structure of loss landscape of deep neural networks
Figure 4 for Embedding Principle: a hierarchical structure of loss landscape of deep neural networks
Viaarxiv icon