Thanks for your amazing work! I have a question about the latent states of S4. S4 is based on SSM, in which x(t) 's latent dynamics are explicitly modeled. S4 uses discretization and convolution to avoid integration of x'(t), which makes it much more efficient. I wonder if you have considered using the latent states for the interpretation of the latent dynamics of the data. If so, in a deep S4 model, the latent states of which layer would be most meaningful for interpretation?
Thanks for your amazing work! I have a question about the latent states of S4. S4 is based on SSM, in which x(t) 's latent dynamics are explicitly modeled. S4 uses discretization and convolution to avoid integration of x'(t), which makes it much more efficient. I wonder if you have considered using the latent states for the interpretation of the latent dynamics of the data. If so, in a deep S4 model, the latent states of which layer would be most meaningful for interpretation?
Thank you so much in advance!