Qedma and HQC² Demonstrate 30–50× Error Mitigation Advantage in Quantum Chemistry Benchmark

Qedma and HQC² Demonstrate 30–50× Error Mitigation Advantage in Quantum Chemistry Benchmark Quantum error mitigation software provider Qedma Quantum Computing and the HQC² research consortium—comprising the University of Copenhagen (KU), Technical University of Denmark (DTU), and University of Southern Denmark (SDU)—have published an experimental benchmark demonstrating a 30× to 50× reduction in energy estimation error for quantum chemistry calculations on superconducting quantum hardware. Presented at Q2B Copenhagen 2026, the study evaluated Qedma’s characterization-based QESEM (Quantum Error Suppression and Error Mitigation) software layer on IBM’s Aachen quantum processor, computing the potential energy surface (PES) of a symmetrically stretched water molecule. The research team mapped a chemically motivated, orbital-optimized variational ansatz—the single-layer perfect-pairing tiled unitary product state (pp-tUPS)—onto an 8-qubit register corresponding to a (4,4) active space in the STO-3G basis set. Unmitigated QPU executions routinely overestimated the ground-state energy by roughly 500 mHa due to cumulative gate infidelities and decoherence. By integrating QESEM’s noise-aware transpilation and unbiased quasi-probabilistic error cancellation (PEC) algorithms, the team brought energy estimation errors systematically down to within 100 mHa, 30 mHa, and under chemical accuracy (~1.5 mHa) relative to statevector references as target precision constraints were tightened. [ Water Molecule (H₂O) Potential Energy Surface Experimental Benchmarks ]Execution & Mitigation TierEnergy Estimation Error (vs. Statevector)Mitigation Factor / Target AccuracyRaw QPU Execution (IBM Aachen)~500 mHa OverestimationBaseline Unmitigated Hardware NoiseQESEM Loose Precision Target (0.1 Ha)~30 mHa to 100 mHa Error Range30× Error Reduction; Moderate Sampling OverheadQESEM Tight Precision Target (0.01 Ha)~1.5 mHa (Chemical Accuracy Horizon)50× Error Reduction; Reaches Target Chemical Accuracy Unlike heuristic error mitigation protocols that fail to provide formal mathematical guarantees, QESEM applies characterization-based, noise-aware transpilation that directly mitigates non-Clifford fractional rotation gates without requiring conversion into longer Clifford gate sequences. This preserves active circuit volume while systematically eliminating systematic noise bias. While the study highlighted the substantial QPU sampling overhead required to achieve sub-millihartree precision on current hardware, it demonstrated that characterization-driven QEM can extract quantitative molecular properties and smooth potential energy surfaces prior to full fault-tolerant hardware deployment. Funded in part by the Israel Innovation Authority under the Eureka Q-CHEMION project and the Novo Nordisk Foundation, the collaboration leverages HQC²’s open-source SlowQuant computational chemistry suite to calculate ground- and excited-state molecular properties on near-term devices. The benchmark follows Qedma’s recent joint demonstrations with IBM and Quantinuum showing quantum advantage in modeling 2D Floquet physics beyond the limits of classical supercomputers. Review the official news announcement on Qedma here, access the technical research paper on arXiv PDF here, read our coverage of Qedma’s QESEM Integration with Quantinuum Hardware here, and examine our previous analysis of IBM and Qedma’s Floquet Physics Quantum Advantage Benchmarks here. September 9, 2026 Mohamed Abdel-Kareem2026-09-09T18:16:20-07:00 Leave A Comment Cancel replyComment Type in the text displayed above Δ This site uses Akismet to reduce spam. Learn how your comment data is processed.