New study probes entanglement in noisy quantum processors

Researchers are discovering a counterintuitive relationship between local measurements and long-range entanglement in quantum processors. The study reveals that sampling from even simple, constant-depth 2D circuits can be computationally difficult, despite being relatively easy to create. This difficulty arises because local measurements can unexpectedly convert short-range entanglement into long-range entanglement, a phenomenon linked to the efficiency of algorithms used to simulate these circuits. Understanding this dynamic is crucial for determining if noisy 2D random-circuit sampling can be efficiently simulated, as the measurement-induced entanglement serves as an indicator of the algorithm. Entanglement’s Role in Quantum Advantage & Classical Simulation Constant-depth 2D circuits, despite being relatively simple to create, present a surprising challenge to classical computation; researchers have long conjectured they are average-case hard to sample from, meaning determining the probability of any given outcome requires immense computational resources. This difficulty isn’t simply a matter of circuit complexity, but stems from how measurements within these circuits manipulate entanglement, transforming localized connections into long-range correlations. The work demonstrates that sampling from even these seemingly straightforward circuits can be computationally intractable, a finding with significant implications for assessing the potential of near-term quantum processors. It is known that when circuit depth scales at least logarithmically with system size, any depolarizing noise of constant rate p greater than zero renders 2D random-circuit sampling amenable to efficient classical simulation. By contrast, in the sub-logarithmic depth regime, efficient classical simulation under depolarizing noise has been established only at large noise rates p greater than p c, via approaches such as trajectory unraveling and percolation-based analyses. Clarifying the fate of measurement-induced entanglement at small, constant noise rate p greater than zero and at sub-logarithmic depths would not only illuminate the interplay among unitary gates, measurements, and noise, but also provide crucial insight into closing the remaining gap in our understanding of the simulation complexity of noisy 2D random-circuit sampling. To quantify measurement-induced entanglement, the team simulated sampling the 2D state column-by-column and considered the operator entanglement S op (t) of the boundary state ρ t generated by the sampling process. They identified a transition from area-law to volume-law behavior at a critical depth of six in the noiseless limit, consistent with previous studies and related to measurement-induced phase transitions. However, the most striking finding concerns the impact of noise. The study reports that “for any constant noise rate p greater than zero, the maximal S op obeys an area-law,” meaning the entanglement remains confined and doesn’t scale with system size. This contrasts with the volume-law scaling observed in noiseless circuits, and suggests that even small amounts of noise can significantly limit the growth of entanglement. Stabilizer generators are distributed almost evenly in the bulk of the boundary state, with their length occurring with a probability proportional to e to the power of negative γ p,T, bounding the number crossing a bipartition. This bounds the number of stabilizer generators crossing a bipartition, underpins the area-law scaling of S op max, and implies exponential decay of the conditional mutual information. This area-law scaling is crucial because it suggests that the MPO-SEBD algorithm, an extension of the space-evolving block decimation algorithm, can efficiently sample from noisy 2D random Clifford circuits.

The team demonstrated the algorithm on noisy 2D Haar-random circuits and measurement-based quantum computing circuits, providing evidence that noise can indeed destroy the volume-law scaling of measurement-induced entanglement in non-Clifford circuits, potentially enabling more efficient classical simulations of these quantum systems. Measurement-Induced Entanglement in 2D Random Circuits Recent work has begun to illuminate the role of measurement-induced entanglement in driving this difficulty, revealing how local measurements can unexpectedly transform short-range entanglement into long-range correlations. Entanglement is widely recognized as a key resource enabling tasks unattainable in a classical world, including unconditionally secure cryptography and computational speed-ups. Generic highly entangled many-body states are believed to be hard to describe and simulate on classical hardware, and creating a large amount of entanglement is therefore a natural ingredient in any attempt to achieve quantum advantage. These findings offer a crucial step toward understanding the limits of classical simulation for noisy 2D random-circuit sampling, and provide a foundation for designing more robust quantum architectures. Noise Suppression of Volume-Law Scaling of MIE Researchers at Google Quantum AI are meticulously mapping the behavior of entanglement in increasingly complex quantum circuits, seeking to understand the limits of classical computation in simulating these systems. Their recent work focuses on measurement-induced entanglement in two-dimensional random Clifford circuits, a model chosen for its relative ease of simulation while still capturing key features of more general quantum systems.

The team’s investigations reveal a surprising interplay between noise and entanglement, challenging previous assumptions about how these circuits scale in complexity. The new research demonstrates that even small amounts of noise fundamentally alter this entanglement structure. To characterize this phenomenon, the team analyzed the spatial distribution of stabilizer generators, key components in understanding entanglement. The implications of this work extend to the development of classical algorithms for simulating quantum circuits. They found that the MPO-SEBD algorithm guarantees efficient sampling throughout the entire sub-logarithmic depth regime T equals o (log N) for arbitrarily small constant noise rate p greater than or equal to one, and at constant depths T equals O (1) for noise rates p greater than or equal to the inverse of log N, within polynomial time. Sub-Logarithmic Depth & Classical Simulability The pursuit of scalable quantum computation hinges on understanding how easily quantum states can be replicated by conventional computers; recent work illuminates a surprising link between noise, entanglement, and the limits of classical simulation in two-dimensional quantum circuits. While creating complex entangled states is considered a prerequisite for quantum advantage, researchers have discovered that even relatively simple circuits can pose significant computational challenges, particularly when subjected to realistic levels of noise. Specifically, the team investigated circuits with a depth that scales sub-logarithmically with system size, meaning their complexity grows slower than the logarithm of the number of qubits. These circuits, despite their constrained structure, are conjectured to be average-case hard to sample from, a counterintuitive finding given the ease with which simpler circuits are typically analyzed. The study focused on 2D random Clifford circuits, chosen for their efficient simulability which allowed for detailed numerical analysis at small noise rates. The MPO-SEBD algorithm guarantees efficient sampling throughout the entire sub-logarithmic depth regime T equals o (log N) for arbitrarily small constant noise rate p greater than or equal to one, and at constant depths T equals O (1) for noise rates p greater than or equal to the inverse of log N, within polynomial time. A key finding concerns the behavior of ‘measurement-induced entanglement,’ a phenomenon where sampling from a portion of the circuit creates highly entangled states on the remaining qubits. However, the introduction of even a constant noise rate fundamentally alters this behavior. This understanding of entanglement distribution has direct implications for classical simulation algorithms. The results suggest that even in the presence of noise, the computational difficulty associated with these circuits may be less severe than previously anticipated, potentially narrowing the path toward demonstrating quantum advantage.

Clifford Circuits Simulate Entanglement Dynamics The assumption that complex quantum states require deep circuits to generate is being challenged by new research into entanglement dynamics, particularly within noisy systems. While building entanglement typically demands substantial circuit depth, studies reveal a surprising resilience of entanglement even in relatively simple, constant-depth two-dimensional circuits, and a counterintuitive role for noise in shaping its distribution. In the noiseless limit, they identified a finite-depth transition from area- to volume-law behavior with a critical depth of six, consistent with previous studies.

Operator Entanglement Scaling with Noise & Depth The surprising resilience of entanglement in noisy quantum circuits challenges conventional wisdom about computational complexity. Analysis of the spatial distribution of stabilizer generators provides insight into this phenomenon. Researchers define two key distributions: the center-location distribution C(x c) and the length distribution D(ℓ). These are defined as the probabilities that a randomly picked stabilizer generator g has center com (g) equal to x c and length len (g) equal to ℓ, respectively. This analysis shows that, as noise increases, the length distribution shifts, indicating a reduction in long-range entanglement. The researchers extended the space-evolving block decimation technique, dubbed MPO-SEBD, demonstrating its ability to efficiently sample from these noisy circuits under certain conditions. The findings have implications for assessing the classical simulability of noisy 2D random-circuit sampling, and provide crucial insight into the interplay among unitary gates, measurements, and noise.

Stabilizer Generator Distribution & Area-Law Behavior Investigations into the behavior of entanglement within quantum circuits are revealing a surprising connection between the distribution of stabilizer generators and the emergence of area-law scaling, even in the presence of noise. Researchers are focusing on how these generators, crucial for characterizing entanglement, are spatially arranged within the boundary state of a two-dimensional circuit undergoing column-by-column sampling. Analysis of this distribution provides insight into the classical simulability of these circuits, a key question in the pursuit of quantum advantage.

The team’s work demonstrates that even relatively simple constant-depth 2D circuits can present computational challenges, defying expectations that shallower circuits would be easier to analyze. This difficulty, they find, stems from the way local measurements can transform entanglement from short-range to long-range, creating complex correlations that are hard to replicate classically. These distributions reveal how the generators are positioned and how their lengths vary across the boundary state. Notably, the researchers discovered a stark difference between noiseless and noisy circuits. This decay is quantified by a coefficient that increases monotonically with noise levels, demonstrating that noise consistently reduces the overall entanglement. The researchers also observed that the mutual information, a measure of correlation between different parts of the system, decreases exponentially with the length of the stabilizer generators. MPO-SEBD Algorithm Enables Efficient Sampling Researchers at Google Quantum AI are refining techniques to assess the complexity of simulating quantum circuits, focusing on how noise impacts the efficiency of certain algorithms. Their recent work centers on the Matrix Product Operator, Space Evolving Block Decimation algorithm, a method for sampling from noisy 2D random Clifford circuits, and demonstrates its potential for efficient simulation under specific conditions.

The team’s investigation addresses a critical question: can the entanglement inherent in these circuits, and the computational difficulty of simulating them, persist even with the introduction of realistic noise? The study characterizes the spatial structure of entanglement generated during the sampling process, using a metric called S op to quantify operator entanglement across a horizontal half-space. They discovered a significant shift in behavior between noiseless and noisy circuits. Further analysis revealed how stabilizer generators, key components in describing the circuit’s entanglement, are distributed within the boundary state. MIE in Haar-Random vs.

Clifford Circuits While generating substantial entanglement is considered vital for quantum advantage, the relationship between entanglement and computational complexity isn’t straightforward, particularly when noise is present. Researchers have long known that constant-depth two-dimensional circuits, despite their relative simplicity, present a challenge for classical simulation, and this difficulty appears linked to measurement-induced entanglement, a phenomenon where local measurements unexpectedly create long-range correlations. They discovered a critical shift in behavior depending on circuit depth. Further analysis revealed how the distribution of stabilizer generators, key components describing circuit entanglement, changes with noise. 👉 More information🗞 Measurement-Induced Entanglement in Noisy 2D Random Circuits✍️ Zhi-Yuan Wei, Jon Nelson, Joel Rajakumar, Esther Cruz, Alexey V. Gorshkov, Michael J. Gullans and Daniel Malz🧠 DOI: http://link.aps.org/doi/10.1103/yfzm-5rr6 Stay currentSee today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals. Tags: