Diff2Dist: Learning Spectrally Distinct Edge Functions, with Applications to Cell Morphology Analysis

Fig. 1. A neural network model learns edge weights which distinguish two classes of graphs. Each row shows the weight values assigned by the network at different times during the training process, from pre-training (far left) to convergence (right). The top row represents a patch of wild-type Arabidopsis cells, and the bottom row represents mutants. The pictured edge weights cause these two categories of graph to have distinct spectra.

[1]  Max Welling,et al.  Semi-Supervised Classification with Graph Convolutional Networks , 2016, ICLR.

[2]  Kunihiko Fukushima,et al.  Neocognitron: A Self-Organizing Neural Network Model for a Mechanism of Visual Pattern Recognition , 1982 .

[3]  The Arabidopsis Genome Initiative.  Analysis of the genome sequence of the flowering plant Arabidopsis thaliana , 2000, Nature.

[4]  J. Tenenbaum,et al.  A global geometric framework for nonlinear dimensionality reduction. , 2000, Science.

[5]  Y. Couder,et al.  Developmental Patterning by Mechanical Signals in Arabidopsis , 2009 .

[6]  David K. Hammond,et al.  Graph diffusion distance: A difference measure for weighted graphs based on the graph Laplacian exponential kernel , 2013, 2013 IEEE Global Conference on Signal and Information Processing.

[7]  Jimmy Ba,et al.  Adam: A Method for Stochastic Optimization , 2014, ICLR.

[8]  Richard B. Nelson,et al.  Simplified calculation of eigenvector derivatives , 1976 .

[9]  Graph diffusion distance: Properties and efficient computation , 2021, PloS one.

[10]  D. Bouchez,et al.  The preprophase band of microtubules controls the robustness of division orientation in plants , 2017, Science.

[11]  K. E. Chu,et al.  Derivatives of Eigenvalues and Eigenvectors of Matrix Functions , 1993, SIAM J. Matrix Anal. Appl..

[12]  Yann LeCun,et al.  Dimensionality Reduction by Learning an Invariant Mapping , 2006, 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06).

[13]  Natalia Gimelshein,et al.  PyTorch: An Imperative Style, High-Performance Deep Learning Library , 2019, NeurIPS.