Picture for Molei Tao

Molei Tao

gDDIM: Generalized denoising diffusion implicit models

Jun 11, 2022
Qinsheng Zhang, Molei Tao, Yongxin Chen

Figure 1 for gDDIM: Generalized denoising diffusion implicit models

Figure 2 for gDDIM: Generalized denoising diffusion implicit models

Figure 3 for gDDIM: Generalized denoising diffusion implicit models

Figure 4 for gDDIM: Generalized denoising diffusion implicit models

* 26 pages, 9 figures, implementation https://github.com/qsh-zh/gDDIM

Access Paper or Ask Questions

Momentum Stiefel Optimizer, with Applications to Suitably-Orthogonal Attention, and Optimal Transport

May 27, 2022
Lingkai Kong, Yuqing Wang, Molei Tao

Figure 1 for Momentum Stiefel Optimizer, with Applications to Suitably-Orthogonal Attention, and Optimal Transport

Figure 2 for Momentum Stiefel Optimizer, with Applications to Suitably-Orthogonal Attention, and Optimal Transport

Figure 3 for Momentum Stiefel Optimizer, with Applications to Suitably-Orthogonal Attention, and Optimal Transport

Figure 4 for Momentum Stiefel Optimizer, with Applications to Suitably-Orthogonal Attention, and Optimal Transport

* Comments are welcome

Access Paper or Ask Questions

The Mirror Langevin Algorithm Converges with Vanishing Bias

Oct 11, 2021
Ruilin Li, Molei Tao, Santosh S. Vempala, Andre Wibisono

Access Paper or Ask Questions

Large Learning Rate Tames Homogeneity: Convergence and Balancing Effect

Oct 07, 2021
Yuqing Wang, Minshuo Chen, Tuo Zhao, Molei Tao

Figure 1 for Large Learning Rate Tames Homogeneity: Convergence and Balancing Effect

Figure 2 for Large Learning Rate Tames Homogeneity: Convergence and Balancing Effect

Figure 3 for Large Learning Rate Tames Homogeneity: Convergence and Balancing Effect

Figure 4 for Large Learning Rate Tames Homogeneity: Convergence and Balancing Effect

* Questions and comments are welcome

Access Paper or Ask Questions

Sqrt(d) Dimension Dependence of Langevin Monte Carlo

Sep 23, 2021
Ruilin Li, Hongyuan Zha, Molei Tao

Figure 1 for Sqrt(d) Dimension Dependence of Langevin Monte Carlo

Figure 2 for Sqrt(d) Dimension Dependence of Langevin Monte Carlo

* v1 submitted on May 28, 2021 (NeurIPS 2021 deadline); v2 added an important reference and discussions

Access Paper or Ask Questions

Mean-Square Analysis with An Application to Optimal Dimension Dependence of Langevin Monte Carlo

Sep 08, 2021
Ruilin Li, Hongyuan Zha, Molei Tao

Figure 1 for Mean-Square Analysis with An Application to Optimal Dimension Dependence of Langevin Monte Carlo

Figure 2 for Mean-Square Analysis with An Application to Optimal Dimension Dependence of Langevin Monte Carlo

* Submitted to NeurIPS 2021 on May 28, 2021 (the submission deadline)

Access Paper or Ask Questions