2024 hackathon Max distance -> 0.07
We do a piecewise linear approximation of tanh function in the interval [0,1).
We divide the interval into 4 equal sub-intervals and approximate the function there. For a precision of 4 for |x>, we can output |x>|p(x)> with 24 qubits with l_{inf} norm cost 0.07.
For a precision of 20 for |x>, the same algorithm can work with 100 qubits. We won the first prize!