Google Quantum AI's Quest for Error-Corrected Quantum Computers

TL;DR

This article provides a comprehensive review of Google Quantum AI's pivotal role in the quantum computing landscape over the past decade, emphasizing their significant strides towards achieving quantum computational supremacy.

Abstract

Quantum computers stand at the forefront of technological innovation, offering exponential computational speed-ups that challenge classical computing capabilities. At the cutting edge of this transformation is Google Quantum AI, a leader in driving forward the development of practical quantum computers. This article provides a comprehensive review of Google Quantum AI's pivotal role in the quantum computing landscape over the past decade, emphasizing their significant strides towards achieving quantum computational supremacy. By exploring their advancements and contributions in quantum hardware, quantum software, error correction, and quantum algorithms, this study highlights the transformative impact of Google Quantum AI's initiatives in shaping the future of quantum computing technology.

Figures (12)

• Figure 1: The roadmap for quantum computing at Google Quantum AI illustrates their commitment to unlocking the ultimate potential of quantum computing through the development of a large-scale, error-corrected computer. The journey is guided by six pivotal milestones. The first milestone, beyond classical, was achieved in 2019, marking a significant advance over classical computing art19. The second milestone, error-corrected qubits (see Section \ref{['errorcorrection']}). Subsequent milestones include building a long-lived logical qubit, creating a logical gate, and engineering scale-up. The final milestone, a large error-corrected quantum computer, represents the ultimate goal of connecting and controlling 1 million qubits, pushing the boundaries of quantum technology to realize meaningful applications.
• Figure 2: Demonstration of quantum computational supremacy using SYC-53. The process comprises four steps: First, a specific quantum circuit is selected. Next, the circuit is executed on the quantum processor. Following this, the processor's fidelity is estimated, and the labor cost is assessed. Finally, the results indicate that quantum computational supremacy has been achieved. Reproduced from Supremacy19.
• Figure 3: Google’s quantum computer achieves a significant milestone by decreasing error rates. Researchers have shown for the first time that increasing the number of qubits can reduce the error rate in quantum calculations GoogleAI2023_1. Image source Google_image. Credit: Google Quantum AI.
• Figure 4: A schematic representation shows a conical surface in the ($x, y, t$) space, encompassing the effective volume $V_{\text{eff}}$ that envelops the tensorial structure of entangling gates affecting the local observable $B(t) = U^\dagger(t) B U(t)$, with $B(0) = B$ indicated by a black dot. The cross-section of this cone, with area $A_{\text{eff}}$, is highlighted where the plane intersects the cone. The cone’s base represents the subset of qubits in the ($x, y$) plane engaged in the operator spreading of $B$ at time $t$, while the perimeter of this base signifies the scrambling front advancing with velocity $v_B$. Reproduced under a creative common license ( https://creativecommons.org/licenses/by/4.0/) from effectiveQV.
• Figure 5: Architectural overview of Google's Bristlecone quantum processor. The sub-lattices of the complete Bristlecone-72 (shown in the bottom right) are arranged in order of increasing complexity for a given depth. Notably, Bristlecone-72 is not more difficult to simulate than Bristlecone-70, as the corner tensors can be contracted with minimal computational expense. Additionally, Bristlecone-64 exhibits similar simulation complexity to Bristlecone-48, while being significantly easier to simulate than Bristlecone-60 GoogleAI36. Google Quantum AI has identified a series of sub-lattices, specifically Bristlecone-24, Bristlecone-30, Bristlecone-40, Bristlecone-48, Bristlecone-60, Bristlecone-64, and Bristlecone-70, arranged from top left to bottom left, all of which present significant challenges for classical simulation while maintaining a low qubit count.