CPC G06N 3/045 (2023.01) [G06N 3/088 (2013.01)] | 20 Claims |
1. A method comprising, by one or more computing systems:
accessing a first data matrix comprising a plurality of row data and a plurality of column data;
providing, to a first generative adversarial network (GAN), a first data input comprising a plurality of row vectors corresponding to the plurality of row data from the first data matrix;
providing, to a second GAN, a second data input comprising a plurality of column vectors corresponding to the plurality of column data from the first data matrix; and
generating a trained a neural-network row generator of the first GAN by, for each of the plurality of row vectors:
generating, by the neural-network row generator of the first GAN, fake row data;
classifying, by a row discriminator of the first GAN, and based on real row-data input, each fake row data as real or fake;
when all fake row data is not classified as real by the row discriminator of the first GAN, then updating one or more weights of the neural-network row generator through backwards propagation; and
when all fake row data is classified as real by the row discriminator of the first GAN, then outputting a set of latent row data based on the real row data and the fake row data generated by the trained neural-network row generator;
generating a trained a neural-network column generator of the second GAN by, for each of the plurality of column vectors:
generating, by the neural-network column generator of the second GAN, fake column data;
classifying, by a column discriminator of the second GAN, and based on real column-data input, each fake column data as real or fake;
when all fake column data is not classified as real by the column discriminator of the second GAN, then updating one or more weights of the neural-network column generator through backwards propagation; and
when all fake column data is classified as real by the column discriminator of the second GAN, then outputting a set of latent column data based on the real column data and the fake column data generated by the trained neural-network column generator; and
generating, by simultaneously co-clustering the plurality of row vectors and the plurality of column vectors by the first GAN and the second GAN, a co-clustered correlation matrix based at least in part on the plurality of row vectors, the plurality of column vectors, the latent row data, and the latent column data, wherein the co-clustered correlation matrix comprises co-clustered associations between the plurality of row data and the plurality of column data of the first data matrix.
|