Second-Quantized Fermionic Operators with Polylogarithmic Qubit and Gate Complexity

We present a method for encoding second-quantized fermionic systems in qubits when the number of fermions is conserved, as in the electronic structure problem. When the number F of fermions is much smaller than the number M of modes, this symmetry reduces the number of information-theoretically required qubits from Θ(M) to O(FlogM). In this limit, our encoding requires O(F2log4M) qubits, while encoded fermionic creation and annihilation operators have cost O(F2log5M) in two-qubit gates. When incorporated into randomized simulation methods, this permits simulating time evolution with only polylogarithmic explicit dependence on M. This is the first second-quantized encoding of fermions in qubits whose costs in qubits and gates are both polylogarithmic in M, which permits studying fermionic systems in the high-accuracy regime of many modes.