Error Mitigation
Squeezing better answers out of noisy hardware without fixing the errors, using extra runs and statistics. The budget alternative to full error correction.
Error mitigation is the art of getting better answers out of a noisy quantum computer without actually fixing the errors. Run the circuit at several amplified noise levels and extrapolate to zero noise; characterize the noise and statistically invert it; average over cleverly randomized variants. No extra qubits required, just extra runs and some statistical nerve.
It's the budget alternative to error correction: correction detects and repairs errors mid-flight but demands enormous qubit overhead, while mitigation accepts the noise and cleans up in post-processing. The catch is that mitigation's costs grow exponentially with circuit size, so it buys a useful-but-bounded window, not a road to scale.
Reading vendor results, the question to ask: were these numbers error-mitigated, and by how much? The technique is legitimate; the fine print is mandatory.