Exploring KANs : A Unique Approach to Function Approximation in AI