Campuses:

Dongmian Zou

  • Industrial Postdoctoral Fellow
Location: 
433 Lind Hall
Email: 
dzou@umn.edu

Ph.D. Applied Mathematics, University of Maryland

IMA Project

Invariances for graph convolutional neural networks

  • generalized the scattering transform to graphs and consequently construct a convolutional neural network on graphs
  • proved such system enjoys invariance to permutations and stability to graph manipulations
  • numerically decomstrated competitive performance on relevant datasets

IMA Industrial Projects at Cargill

Importance analysis for components of rapeseed oil prices

  • studied relative importance of commodities that impact rapeseed oil prices and used it for price prediction
  • developed an iterative Lasso method for stable representation of components
  • developed a multiresolution method for generating relative importance according to decision tree evaluation process

Correlation analysis between soybean prices in different markets

  • studied broker behaviors that affect soybean product prices at DCE and CME
  • applied dynamical time warping and symbolic aggregate to classifying hedgers
  • applied wavelet and coherence analysis to comparing DCE and CME prices

Related Publications

  1. D. Zou, G. Lerman, Graph Convolutional Neural Networks via Scattering, sub- mitted. arXiv: 1804.00099.
  2. R. Balan, M. Singh and D. Zou, On Lipschitz bound of general convolutional neural networks, submitted. axXiv: 1808.01415.
  3. R. Balan, M. Singh and D. Zou, Lipschitz Properties for Deep Convolutional Networks, Contemp. Math. 706 (2018) 129–151.

Related Talks and Events

  1. AMS Sectional Meeting, San Francisco (upcoming), October 2018
  2. IMA Postdoc Seminar. "Phaseless reconstruction: A frame theoretical approach." Minneapolis, Minnesota. October 27 and November 3, 2017
  3. IMA Data Science Seminar. "Lipschitz bounds for general convolutional neural networks." Minneapolis, Minnesota. September 2017
  4. ICERM-HKUST Joint Workshop on Phase Retrieval. "Lipschitz extension for phase retrieval." Rhode Island. June 2017 
  5. Reviewer for Measurement: Journal of the International Measurement Confederation. May 2018–Present 
  6. Helper for IMA Math Career Workshop for Iowa State University. May 2018
  7. Participant at February Fourier Talks. University of Maryland. February 2018
  8. Participant at IPAM workshop on New Deep Learning Techniques. UCLA. February 2018