Quantum noise tolerance unlocks IQM’s new optimization technique

Instead of demanding more from increasingly complex quantum hardware, IQM has demonstrated a technique for extracting greater value from existing quantum outputs. The company’s researchers developed a method called quantum-informed surrogate sampling (QISS) which analyzes statistical fingerprints, averages describing how variables align, produced during Quantum Approximate Optimization Algorithm (QAOA) runs.This illustrates what Production Quantum looks like in practice: not waiting for bigger, cleaner hardware, but getting more useful answers out of the machines already running, the team reports in a recent paper. By applying classical Markov chain Monte Carlo sampling to these fingerprints, IQM achieved solution quality comparable to established classical heuristics on its 54-qubit Emerald processor, even with existing noise.IQM has demonstrated a technique for improving the performance of quantum optimization algorithms by leveraging classical computing to analyze data already produced by its 54-qubit processor, Emerald. The core innovation lies in how IQM utilizes information typically discarded after a QAOA run. These represent averages describing the alignment of variables within the optimization problem and are relatively unaffected by the noise inherent in current quantum systems.While conventional methods focus solely on the final candidate solutions output by QAOA, QISS treats these fingerprints as a valuable resource.

The team then constructs a classical model, a structured map, from these fingerprints and employs Markov chain Monte Carlo sampling to generate improved solutions. This classical analysis requires no additional quantum measurements or parameter tuning, representing a significant efficiency gain, and extends beyond immediate performance improvements.Because the quantum measurements are already taken, and the classical post-processing is computationally inexpensive, the method’s reliance on averaged data makes it remarkably resilient to noise; small errors in measurement gently alter the results rather than causing catastrophic failures. Testing on the Emerald processor demonstrated that QISS recovered performance levels approaching those of an ideal, noiseless quantum computer. This noise tolerance is particularly important given recent estimates suggesting that achieving a substantial advantage over classical solvers may require upwards of a million physical qubits, even with error correction.Better post-processing lowers the bar for quantum hardware, both now and in the future. IQM tested the QISS method on two challenging optimization problems, MaxCut and Maximum Independent Set, across varying problem sizes.The technique was not merely a theoretical exercise; it was “developed and verified on our own 54-qubit processor, Emerald, not as a hypothetical benchmark but as a working part of the stack.” This full-stack approach, where the same team manages both the quantum hardware and the classical post-processing layer, provides a distinct advantage. The ability to squeeze more performance from existing quantum hardware has significant implications for the near-term viability of quantum optimization, offering a pathway to enhance results without waiting for substantial advancements in quantum hardware. Elisabeth Wybo, Team Lead Optimization at IQM, emphasizes the practical benefits of this strategy.Wybo, who holds a PhD in quantum many-body physics from the Technical University of Munich and has served as Team Leader, Optimisation since October 2024, highlights that QISS represents a concrete example of how a full-stack company can deliver tangible improvements in quantum computing performance. This strategy positions IQM to capitalize on the evolving quantum landscape by maximizing the utility of available resources and accelerating the path toward practical quantum solutions. Source: https://iqm.tech/blog/getting-more-out-of-quantum-optimization/ See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.Ivy Delaney has been working with neural networks and machine learning since the mid-nineties, back when a couple of hidden layers and a long afternoon of training counted as ambitious. She has watched the field go from academic curiosity to the thing quietly running underneath everything, and she brings that long view to quantum computing.

For Quantum Zeitgeist she covers the ground where the two fields meet. That means quantum machine learning and the variational algorithms it leans on, and it also means the less glamorous but more interesting story of classical machine learning already doing real work inside quantum machines, decoding error-correcting codes, calibrating noisy hardware and learning the error models that simulators depend on. She writes about the hardware those algorithms have to run on too, and about the post-quantum cryptography scramble that the same hardware has set off. Her stories typically start with the paper, whether that is peer-reviewed work, conference proceedings or an arXiv preprint, with the source linked so you can hold a claim up against the research it came from. She is unimpressed by benchmarks that will not say what they beat, and by demonstrations that only work in the press release.