· 智源学者

博士,1986年1月出生,主要研究领域为机器学习,运筹学和图像处理。研究内容包括深度学习模型的数学原理,分布式随机优化算法,高维数据的统计模型。博士期间提出了新的并行随机优化算法,显著加速和改善了深度学习模型的训练;博士后期间提出了相位调和变换,并以此解释了卷积神经网络中的激活函数ReLu的作用。

 

代表性成果

 

在机器学习,应用数学和相关交叉学科的顶级期刊发表。

  1. To submit to Applied and Computational Harmonic Analysis: Wavelet Phase Harmonic Covariance Models of Stationary Processes, with Stéphane Mallat.
  2. To submit to JMLR software. Kymatio: Wavelet Scattering Transforms in Python, with Mathieu Andreux et al.
  3. Submitted to Information and Inference of IMA: Phase Harmonics and Correlation Invariants in Convolutional Neural Networks, with Stéphane Mallatet al.
  4. Submitted to Astronomy & Astrophysics: The reduced wavelet scattering transform - a general statistical description for non-linear physical processes,with Erwan Allys et al.
  5. Submitted to ICASSP: Moment-matching audio synthesis phase harmonics, with Gaspar Rochette et al.
  6. Brochard, A, Blaszczyszyn, B, Mallat, S, and Zhang, S.Statistical learning of geometric characteristics of wireless networks. IEEE INFOCOM 2019.
  7. Sixin Zhang, Anna Choromanska, Yann LeCun. Deep learning with Elastic Averaging SGD. in the Neural Information Processing Systems Conference (NIPS). 2015.
  8. Li Wan, Matthew Zeiler, Sixin Zhang, Yann LeCun, Rob Fergus. Regularization of Neural Networks using DropConnect. ICML. 2013.
  9. Tom Schaul, Sixin Zhang, Yann LeCun. No more pesky learning rates. ICML. 2013.

拟研究项目名称:数据的非线性表示和随机过程模型

All Posts
×

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!

OK