Multiscale invariant representations for learning on high dimensional data
Speaker: Matthew Hirn, Michigan State University
Abstract: : Convolutional neural networks (ConvNets) have revolutionized our approach to learning tasks for high dimensional signal data by processing a signal through a cascade of learned convolution operators and nonlinear operations, which successively extract information from the signal that can be used for downstream tasks such as classification or regression. Motivated by the successes of ConvNets and inspired by multiscale problems in chemistry, physics, and biology, in this talk I will introduce the wavelet scattering transform, which on the one hand can be viewed as a simplified mathematical model for ConvNets, but on the other hand incorporates unique algorithmic design choices. The wavelet scattering transform replaces the learned filters of ConvNets with predefined wavelets, which are multiscale, oscillating waveforms with zero average, and computes a cascade of alternating wavelet transforms and nonlinear operators. Unlike ConvNets, which are task driven, a wavelet scattering transform is motivated by invariance and stability properties inherent in the data. Here we will focus on problems at the interface of invariant representation learning and the chemical, physical, and biological sciences, such as machine learning for many particle systems and materials science; multi reference alignment inverse problems and super resolution; and the synthesis of random processes that occur naturally (e.g., textures and turbulence).
Room Reservation Information
Room Number: 114 McAllister
Time: 12:20pm - 1:30pm