absorb.md

Travis Humble

Chronological feed of everything captured from Travis Humble.

Bootstrapped Subcircuit Characterization Enables Accurate Modeling of Noisy Quantum Circuits

A test-driven method decomposes NISQ application circuits into subcircuits for individual experimental characterization, yielding a composite model of the full noisy circuit. This approach predicts outcomes for related programs with high accuracy, measured by total variation distance on superconducting transmon devices running GHZ-state preparation and Bernstein-Vazirani algorithms. Models balance computational efficiency against tunable predictive accuracy without full-circuit simulations.

XACC: Hardware-Agnostic System Infrastructure for Integrated Quantum-Classical Programming

XACC provides a service-oriented, system-level software infrastructure for heterogeneous quantum-classical computing, shifting from high-level REST APIs to low-level co-processor models. It exposes modular interfaces for quantum programming, compilation, and execution, remaining hardware-agnostic for both NISQ and future architectures. The framework enables tight integration of quantum and classical workflows, demonstrated through paradigmatic tasks, and supports development of compilers and runtimes.

Hybrid Quantum-Classical Algorithms Excel for Large-Scale Discrete-Continuous Optimization

Hybrid models integrate deterministic classical algorithms with quantum computing to tackle combinatorial complexity in large-scale mixed-integer programming. Applied to molecular conformation, job-shop scheduling, manufacturing cell formation, and vehicle routing problems, these approaches yield superior solution quality and computation time. Results demonstrate that leveraging quantum features complements classical methods for computationally challenging instances.

Quantum Chemistry Benchmark Reveals High Noise in NISQ Devices but Achieves Chemical Accuracy with Error Mitigation

Researchers developed a VQE-based quantum chemistry benchmark incorporating active space reduction, reduced unitary coupled cluster ansatz, and McWeeny density purification for NISQ devices. Simulations of alkali metal hydrides (NaH, KH, RbH) on IBM Tokyo (20 qubits) and Rigetti Aspen (16 qubits) highlight characteristic high noise in superconducting hardware. Post-processing error mitigation, particularly density purification, dramatically improves ground-state energy accuracy, enabling chemical accuracy for specific settings via cloud access.

Quantum Annealing Solves Nurse Scheduling Satisfactorily, Enhanced by Reverse Annealing

Quantum annealing on D-Wave 2000Q solves Nurse Scheduling Problem (NSP) instances by mapping to an Ising-type Hamiltonian, yielding solutions that satisfy hard constraints. Empirical tests show the method produces diverse, practical solutions. Reverse annealing significantly improves solution quality by refining initial results through a second annealing pass.

Dynamic Graphs Enable Universal Quantum Logic via Continuous-Time Quantum Walks

Continuous-time quantum walks (CTQWs) extend from static to dynamic graphs, where a sequence of graphs drives the walk's free evolution. Perfect state transfer in these walks designs dynamic graphs implementing a universal set of quantum logic gates, demonstrated for a complete logical basis. Numerical simulations validate implementations for quantum teleportation and addition circuits. Realization is feasible using actively controlled quantum optical waveguides.

Dynamic Quantum Search Enables Iterative Function Maximization with Updating Constraints

Presents an iterative quantum algorithm for maximizing a function in dynamic models where prior search results refine the acceptable input set. Builds on quantum search with a dynamic oracle that marks items based on updated constraints. Demonstrates correctness via numerical simulations of quantum circuits for the Knapsack problem using explicit arithmetic oracles and comparators up to 30 qubits.

Quantum Annealing Enables Direct Solving of Polynomial Equation Systems

Researchers introduce a direct quantum annealing method to solve general polynomial equation systems, bypassing iterative solvers' variable convergence tied to condition numbers. Validated on second-order polynomials using a commercial annealer, it applies to linear regression and scales with problem size, condition number, and precision. An iterative annealing variant achieves 10^{-8} tolerance for linear systems.