AI/ML+Physics Part 3: Designing an Architecture - presented by Prof. Steve Brunton

AI/ML+Physics Part 3: Designing an Architecture

Prof. Steve Brunton

Prof. Steve Brunton
Slide at 24:03
RESIDUAL NETWORK (RESNET) Xk+1 Euler Integrator! Xk+1 = xk+f(xk) ResNet Block 3x3 CONV BATCH RELU 3x3 CONV BATCH RELU STRIDE=1 NORM STRIDE=1 NORM PAD=1 PAD=1 He, Zhang, Ren, Sun, CVPR, 2015
1
References
  • 1.
    https://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/He_Deep_Residual_Learning_CVPR_2016_paper.pdf
Share slide
Summary (AI generated)

This architecture promotes the idea of time stepping forward, making it ideal for time series data and dynamical systems. The unit architecture is highly effective for super resolution image segmentation.