Abstract: According to one aspect of the disclosure, a system for performing quantum computations includes: a first environment for being maintained at an ambient temperature, having a classical computing processor; and a second environment for being maintained at a cryogenic temperature, having a plurality of qubits and a multiplexing circuit coupled to the plurality of qubits, each of the plurality of qubits having a gap transition frequency; wherein the classical computing processor is configured to change the states of each of the plurality of qubits by controlling the multiplexing circuit to couple particular ones of the plurality of qubits to a pulsed baseband waveform.

Abstract: Techniques for machine learning assisted qubit state readout are disclosed. A system a set of training data that describes states of multiple qubits, and trains a neural network to determine qubit states based on the set of training data. The system obtains one or more unlabeled qubit signals, and determines one or more states corresponding to the unlabeled qubit signal(s), using the neural network. The unlabeled qubit signal(s) may include one or more multiplexed qubit signals, and the state(s) corresponding to the unlabeled qubit signal(s) may include one or more multi-qubit states based on the multiplexed qubit signal(s).

Abstract: A closed-loop feedback system and method of active noise cancellation to maintain a desired operating frequency of a qubit during a quantum computation, even when that frequency is relatively sensitive to flux noise. A series of Ramsey experiments is performed on the qubit to estimate an offset between its actual and desired operating frequencies, and the error is accumulated. After the probing is complete, the accumulated error is supplied to an arbitrary waveform generator that produces a magnetic flux that is coupled to the qubit, thereby tuning the qubit and actively controlling its operating frequency. Having corrected the operating frequency of the qubit and extended its coherence time, the quantum state of the qubit is allowed to evolve according to the computation.

Abstract: A quantum circuit called a “qumon” is provided to cancel unwanted ZZ interaction in a superconducting qubit architecture. The qumon qubit has a high coherence, and a positive anharmonicity that may be tuned to cancel the negative anharmonicity in a coupled qubit, such as a transmon qubit. The qumon has three parallel branches, in which are a shunt capacitor; a Josephson junction having weighted energy level and capacitance; and several Josephson junctions in series. The weight is chosen to provide the desired anharmonicity, and the transverse flux noise and transverse charge noise each decrease in proportion to the number of the Josephson junctions in series. Because unwanted ZZ interactions are canceled, qumon qubits and transmon qubits may be capacitively coupled in an alternating pattern to provide a surface code in which these interactions are canceled in an extensible way.

Abstract: According to some embodiments, a method can identify and discriminate contributions from one or more noise sources using the multi-level structure of a quantum system with three or more levels. The method can include: preparing the quantum system in a predetermined state; applying one or more control signals to the quantum system; measuring values of one or more observables of the quantum system that quantify the quantum system's response to the noise sources and the one or more applied control signals; extracting noise spectra information associated with the noise sources from the measured values; and identifying contributions from the one or more noise sources based on the noise spectra information.

Abstract: A superconducting qubit is manufactured by stacking up atomically-thin, crystalline monolayers to form a heterostructure held together by van der Waals forces. Two sheets of superconducting material are separated by a third, thin sheet of dielectric to provide both a parallel plate shunting capacitor and a Josephson tunneling barrier. The superconducting material may be a transition metal dichalcogenide (TMD), such as niobium disilicate, and the dielectric may be hexagonal boron nitride. The qubit is etched, or material otherwise removed, to form a magnetic flux loop for tuning. The heterostructure may be protected by adhering additional layers of the dielectric or other insulator on its top and bottom. For readout, the qubit may be coupled to an external resonator, or the resonator may be integral with one of the sheets of superconducting material.

Abstract: A superconducting qubit is manufactured by stacking up atomically-thin, crystalline monolayers to form a heterostructure held together by van der Waals forces. Two sheets of superconducting material are separated by a third, thin sheet of dielectric to provide both a parallel plate shunting capacitor and a Josephson tunneling barrier. The superconducting material may be a transition metal dichalcogenide (TMD), such as niobium disilicate, and the dielectric may be hexagonal boron nitride. The qubit is etched, or material otherwise removed, to form a magnetic flux loop for tuning. The heterostructure may be protected by adhering additional layers of the dielectric or other insulator on its top and bottom. For readout, the qubit may be coupled to an external resonator, or the resonator may be integral with one of the sheets of superconducting material.

Abstract: According to some embodiments, a method can identify and discriminate contributions from one or more noise sources using the multi-level structure of a quantum system with three or more levels. The method can include: preparing the quantum system in a predetermined state; applying one or more control signals to the quantum system; measuring values of one or more observables of the quantum system that quantify the quantum system's response to the noise sources and the one or more applied control signals; extracting noise spectra information associated with the noise sources from the measured values; and identifying contributions from the one or more noise sources based on the noise spectra information.

Abstract: A quantum circuit called a “qumon” is provided to cancel unwanted ZZ interaction in a superconducting qubit architecture. The qumon qubit has a high coherence, and a positive anharmonicity that may be tuned to cancel the negative anharmonicity in a coupled qubit, such as a transmon qubit. The qumon has three parallel branches, in which are a shunt capacitor; a Josephson junction having weighted energy level and capacitance; and several Josephson junctions in series. The weight is chosen to provide the desired anharmonicity, and the transverse flux noise and transverse charge noise each decrease in proportion to the number of the Josephson junctions in series. Because unwanted ZZ interactions are canceled, qumon qubits and transmon qubits may be capacitively coupled in an alternating pattern to provide a surface code in which these interactions are canceled in an extensible way.

Abstract: Techniques for machine learning assisted qubit state readout are disclosed. A system a set of training data that describes states of multiple qubits, and trains a neural network to determine qubit states based on the set of training data. The system obtains one or more unlabeled qubit signals, and determines one or more states corresponding to the unlabeled qubit signal(s), using the neural network. The unlabeled qubit signal(s) may include one or more multiplexed qubit signals, and the state(s) corresponding to the unlabeled qubit signal(s) may include one or more multi-qubit states based on the multiplexed qubit signal(s).