pykan
pykan copied to clipboard
KindXiaoming
How exactly KAN can better interpreted than MLP?
Did you at least read the paper and the documentation of this package? One of the fundamental things pykan is able to do is symbolic transformation, i.e. you are able of expliciting the model formula itself. Every activation is a (smooth) function itself, easily plottable and applicable. I could continue, but I think I already gave too much explanations.
Yes you did I guess .
I read the paper twice and read the documentation too .
I could not connect the idea . Thanks for explaining it to me.
On Mon, May 13, 2024 at 10:27 AM Alessandro Flati @.***> wrote:
Did you at least read the paper and the documentation of this package? One of the fundamental things pykan is able to do is symbolic transformation, i.e. you are able of expliciting the model formula itself. Every activation is a (smooth) function itself, easily plottable and applicable. I could continue, but I think I already gave too much explanations.
— Reply to this email directly, view it on GitHub https://github.com/KindXiaoming/pykan/issues/175#issuecomment-2106655556, or unsubscribe https://github.com/notifications/unsubscribe-auth/AQ56RSWGD5QJVZGCDAANSUTZCBB2DAVCNFSM6AAAAABHTIXK5GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMBWGY2TKNJVGY . You are receiving this because you authored the thread.Message ID: @.***>
Also , can I some how help in making it super easy to understand . Let me know what you think?
On Mon, May 13, 2024 at 10:31 AM Mandar Gite @.***> wrote:
Yes you did I guess .
I read the paper twice and read the documentation too .
I could not connect the idea . Thanks for explaining it to me.
On Mon, May 13, 2024 at 10:27 AM Alessandro Flati < @.***> wrote:
Did you at least read the paper and the documentation of this package? One of the fundamental things pykan is able to do is symbolic transformation, i.e. you are able of expliciting the model formula itself. Every activation is a (smooth) function itself, easily plottable and applicable. I could continue, but I think I already gave too much explanations.
— Reply to this email directly, view it on GitHub https://github.com/KindXiaoming/pykan/issues/175#issuecomment-2106655556, or unsubscribe https://github.com/notifications/unsubscribe-auth/AQ56RSWGD5QJVZGCDAANSUTZCBB2DAVCNFSM6AAAAABHTIXK5GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDCMBWGY2TKNJVGY . You are receiving this because you authored the thread.Message ID: @.***>