The (semi) automatic simplification algorithm provided in the paper for KANs seem, to me, like they're solving a similar problem to https://arxiv.org/pdf/2112.0

The (semi) automatic simplification algorithm provided in the paper for KANs seem, to me, like they're solving a similar problem to https://arxiv.org/pdf/2112.04035 https://arxiv.org/pdf/2112.04035, but with the additional constraint of forward functional interpretability as the goal instead of just a generalized abstraction compressor.