Questions on other usecases

[BradKML]

Are there ways of making tutorials for the following topics (in order of complexity):

• Classification tasks (rather than just regression) https://github.com/team-daniel/KAN
• Recommendation Algorithms (RS) https://github.com/TianyuanYang/KAN4Rec
• Time Series and Temporality (TS) https://github.com/remigenet/TKAN https://github.com/MSD-IRIMAS/Simple-KAN-4-Time-Series
• Reinforcement Learning (RL) https://github.com/riiswa/kanrl
• Embedding and Language Models (LLM) https://github.com/AdityaNG/kan-gpt https://github.com/CG80499/KAN-GPT-2
• Grpah Neural Network (GNN) type applications https://github.com/WillHua127/GraphKAN-Graph-Kolmogorov-Arnold-Networks

[pg2455]

Hey Brad,

I am trying to keep the KAN-Tutorial repo focused on the foundational aspects of KAN to understand what powers its properties. Thus, the notebooks have been focused mostly on extremely simple use cases involving one or two neurons.

I think if some smaller toy problems inspiring the above KAN architectures can be designed, such tutorials can be designed. The links that you suggested are already a good start to inspire a toy problem.

However, I will have to think about what extra knowledge can those tutorials add.

If you have suggestions or examples or a repo you are working on, please feel free to post them for review.