You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I want to contribute an interesting real-world example of using KAN to obtain a symbolic function.
In ocean science, a 75-term expression is used to calculate "density" by using "temperature" and "salinity". I tried to use KAN to find much simpler expression with reasonable accuracy and got good results. The activation shapes also intuitively show that the changes in temperature have more impact on the changes in density compared to salinity.
Very cool! I wonder what that 75-term expression looks like? And how accurate is it? Maybe you can try deeper/wider KANs to get more accuracy but less simplicity. Would be cool to plot Pareto Frontiers trading off between accuracy and simplicity.
Very cool! I wonder what that 75-term expression looks like? And how accurate is it? Maybe you can try deeper/wider KANs to get more accuracy but less simplicity. Would be cool to plot Pareto Frontiers trading off between accuracy and simplicity.
Very interesting work!
I want to contribute an interesting real-world example of using KAN to obtain a symbolic function.
In ocean science, a 75-term expression is used to calculate "density" by using "temperature" and "salinity". I tried to use KAN to find much simpler expression with reasonable accuracy and got good results. The activation shapes also intuitively show that the changes in temperature have more impact on the changes in density compared to salinity.
Example:
https://github.com/obisidianouc/Kan_example.git
happy to discuss!
The text was updated successfully, but these errors were encountered: