Two papers authored by CM3 group members and collaborators are among 51 articles listed in the SIAM Journals Top Three Most Cited Papers since 2018 in the 17 journals published by SIAM.
Recently, this list was published in SIAM New on 11/23/2020. It included a paper on the mathematics of deep learning by the CM3 group, New Error Bounds for Deep ReLU Networks Using Sparse Grids, published in the inaugural issue of the SIAM Journal on Mathematics of Data Science. This 2019 study provided a sparse grid based design of deep neural networks (DNN) and showed that such network structure could help lessen the so-called curse of dimensionality. Another paper on the list, a joint work of Prof.Du and his collaborators, is Maximum Principle Preserving Exponential Time Differencing Schemes for the Nonlocal Allen--Cahn Equation, published in SIAM Journal on Numerical Analysis. The work also motivated further research on time-stepping schemes for more general evolution equations that can preserve physically meaning bounds. A subsequent study by the team will be published in SIAM Review in 2021.
SIAM has given free online access to all the most cited papers on its website. The offer is good until the end of 2020.
