Photo credit: www.sciencedaily.com
Quantum computing stands at the frontier of technological advancement, promising to tackle intricate issues beyond the reach of even the most sophisticated classical supercomputers.
Just as traditional computers consist of interrelated parts, like memory chips and CPUs on a motherboard, quantum computers require effective communication of quantum information across multiple processors. Current interconnection architectures for superconducting quantum processors operate on a “point-to-point” basis, necessitating numerous transfers between network nodes and resulting in cumulative error rates.
In a significant advancement, researchers at MIT have unveiled a novel interconnect device that enables scalable, “all-to-all” communication. This breakthrough allows all superconducting quantum processors within a network to communicate directly, bypassing the limitations of previous architectures.
The research team successfully established a network comprising two quantum processors and utilized their interconnect to facilitate the transfer of microwave photons between them in a user-defined direction. These photons, as carriers of quantum information, play a crucial role in the functionality of quantum systems.
The core of this innovative device is a superconducting waveguide that transports photons between processors and can be configured to connect any number of modules, thus supporting efficient information transmission across a scalable network.
Moreover, the researchers demonstrated the concept of remote entanglement—an essential correlation between disconnected quantum processors. This achievement is a pivotal step toward creating a robust, distributed network of quantum processors.
“Looking ahead, a quantum computer will likely require both local and nonlocal interconnects. While local interconnects are inherent to arrays of superconducting qubits, our design facilitates enhanced nonlocal connections. We can direct photons at various frequencies and times, offering increased flexibility and throughput for our network,” stated Aziza Almanakly, a graduate student in electrical engineering and computer science at MIT and the lead author of the related publication.
Co-authors included Beatriz Yankelevich, a graduate student from the EQuS Group; William D. Oliver, an MIT professor and Lincoln Laboratory Fellow; and various other contributors from MIT and Lincoln Laboratory. This work is set to be featured in Nature Physics.
A scalable architecture
Building on a previous module that allowed the bidirectional transmission of microwave photons via a waveguide, the researchers expanded this concept by linking two modules to the waveguide for photon emission and absorption in a controlled manner.
Each module consists of four qubits, which act as an interface between the waveguide and the broader quantum processors. These qubits facilitate the emission and absorption of photons and relay that information to nearby data qubits for storage.
The researchers employed a sequence of microwave pulses to energize a qubit, prompting photon emission. By carefully manipulating the phase of these pulses, they achieved a quantum interference effect that directed the photon along the waveguide in either direction. By reversing the pulses in time, a qubit from a distant module could absorb the photon regardless of the distance involved.
“By effectively ‘pitching and catching’ photons, we forge a ‘quantum interconnect’ between processors that are not physically adjacent, and this is crucial for establishing remote entanglement,” Oliver elucidated.
“Creating remote entanglement is a fundamental step in constructing a large-scale quantum processor from more modest modules. Once a photon is emitted, the correlation remains between two distant qubits, allowing us to exploit these connections for parallel operations even when they are no longer linked,” added Yankelevich.
However, simple photon transfer between modules is not sufficient to achieve remote entanglement. The research demands meticulous preparation of both the photons and the qubits to enable the modules to effectively “share” the photon after the protocol concludes.
Generating entanglement
The team accomplished this by ceasing the photon emission pulses midway through their duration, resulting in what could be understood as a shared state of the photon. This technique allowed the recipient module to absorb the ‘half-photon,’ thereby establishing entanglement between the two modules.
Throughout the photon’s journey, various physical factors, such as joints and wire bonds within the waveguide, could distort it and impede efficiency during the absorption phase of the receiving module. Consequently, to achieve remote entanglement with sufficient fidelity, the researchers focused on maximizing the absorption rate of the photon.
“One of our primary challenges was tailoring the photon to enhance its absorption efficiency,” commented Almanakly.
To tackle this, the team harnessed a reinforcement learning algorithm to predict how the photon would be distorted during propagation. They could then pre-condition the photon to optimize its shape before transmission.
Upon executing this refined absorption approach, they successfully attained a photon absorption efficiency exceeding 60 percent. This level of efficiency is a remarkable indicator that the resulting state at the conclusion of the protocol is indeed entangled, marking a significant accomplishment in their research.
“This architecture enables the formation of a network with comprehensive connectivity. Multiple modules can be aligned along the same bus, facilitating remote entanglement among any selected pair,” Yankelevich remarked.
Looking to the future, the researchers aim to further enhance absorption efficiency by optimizing the photon’s transmission path—potentially through 3D module integration—while also expediting the protocol to minimize error rates.
“In principle, our method for generating remote entanglement can be adapted for various quantum computer architectures and larger quantum internet systems,” added Almanakly.
This research received support from the U.S. Army Research Office, the AWS Center for Quantum Computing, and the U.S. Air Force Office of Scientific Research.
Source
www.sciencedaily.com