Renormalization-group Prepares Matrix Product States on up to 80 Qubits, Enabling Shallower Circuits for Quantum Systems
Summarize this article with:
Preparing complex, many-body entangled states across numerous qubits represents a significant hurdle in quantum computing, but researchers are now demonstrating substantial progress. Moritz Scheer, Alberto Baiardi, and Elisa Bäumer Marty, all from IBM Quantum, alongside Zhi-Yuan Wei and Daniel Malz, present a method for generating matrix product states using an algorithm rooted in renormalization-group techniques. Their work showcases the preparation of these states on superconducting hardware with systems scaling up to 80 qubits, and reveals that this approach creates circuits with significantly reduced depth compared to traditional methods. This shallower circuit architecture not only improves resilience to noise, but also demonstrably outperforms sequential preparation techniques for larger systems, paving the way for the creation and study of complex quantum states, including those exhibiting symmetry-protected topological order, beyond previously accessible scales. These MPS, representing many-body entangled states, are crucial for simulating complex quantum systems and understanding their behavior.
The team successfully prepared states exhibiting a phase transition between a symmetry-protected topological phase and a trivial phase, scaling systems to up to 80 qubits, the largest demonstration to date of preparing states in this ordered phase away from a fixed point. This new method offers a significant advantage over traditional sequential approaches, achieving exponentially shallower circuit depths as system size increases. This reduced circuit depth enhances resilience to noise, a critical factor for practical quantum computation. Experiments reveal that the reduced circuit depth consistently outperforms sequential circuits for larger systems, demonstrating a clear advantage on currently available hardware. Measurements of string-order-like local expectation values and energy densities confirm the superior scaling of the new protocol with increasing system size, indicating its potential for tackling more complex quantum simulations. Future research will focus on exploring dynamic circuits to further optimize performance and comparing this approach to alternative methods, potentially expanding its applicability to a wider range of quantum systems. Scalable Preparation of Entangled States with Superconducting Qubits Researchers have successfully demonstrated the preparation of matrix product states, complex many-body entangled states, using a novel renormalization-group-based algorithm on superconducting quantum hardware. This approach enables the creation of states with long-range correlations and represents a significant advancement in quantum computing.
The team prepared states exhibiting a phase transition between a symmetry-protected topological (SPT) and a trivial phase for systems containing up to 80 qubits, the largest demonstration to date of preparing states in an SPT ordered phase away from a fixed point. The RG-based protocol delivers a substantial reduction in circuit depth compared to sequential state preparation methods, particularly for larger systems. This reduced circuit depth enhances resilience to noise, a critical factor for practical quantum computation. Measurements of string-order-like local expectation values and energy densities confirm that the RG-based circuits outperform sequential circuits as system size increases, demonstrating a practical advantage on currently available quantum hardware. The core of this achievement lies in the RG-based preparation protocol, which involves iteratively blocking and decomposing the target MPS. This process converges towards a fixed point, allowing for the creation of entangled pairs in constant depth. Strategically choosing an appropriate MPS gauge further reduces the required circuit depth, enhancing efficiency. 👉 More information 🗞 Renormalization-group-based preparation of matrix product states on up to 80 qubits 🧠 ArXiv: https://arxiv.org/abs/2510.24681 Tags: