Training an autoencoder to map the observed oxide compounds and metallic ppm values into latent spaces for improved resource estimation accuracy.
