Quantum Noise Can Boost Computing Power, Research Suggests

J. Montes of the Technical University of Madrid and colleagues have discovered that noise can unexpectedly benefit quantum computation. A geometric mechanism explains how non-unital noise, when combined with standard quantum gates, accelerates the creation of diverse final states. The findings reveal that specific noise induces an expansion of the space of possible quantum states, effectively enriching computational dynamics. Using a single qubit model and the G3 universal gate set, the team analytically determined conditions under which this expansion occurs, suggesting noise could be harnessed as a resource to improve future quantum algorithms.

Noise-induced state space expansion enables faster randomisation in quantum computation

Error rates fell below 1/2 when amplitude damping was combined with the G3 universal gate set, a threshold unattainable in noiseless systems. This improvement represents a fundamental shift in understanding. Achieving Haar-like distributions, indicative of maximal randomness, previously demanded exponentially increasing computational resources, but noise now accelerates this process. The noise induces an effective expansion of the ‘manifold of pure states’, the space defining all possible quantum states, enriching the computational field.

Amplitude damping, a specific type of quantum noise mimicking energy loss, combined with the G3 set of quantum gates, a standard toolkit for building quantum computers, facilitated a faster approach to Haar-like distributions of quantum states. This geometric mechanism, driven by the parameter γ representing the probability of relaxation, allows for quicker exploration of the state space and potentially unlocks new avenues for quantum algorithm design. Renormalization, a technique projecting mixed states back onto pure states, revealed an effective expansion of the space defining all possible quantum states, observed using a single qubit model.

An analytical derivation of a local area expansion factor identified a threshold beyond which this expansion occurs, demonstrating that noise accelerates the creation of maximal randomness, previously requiring exponentially increasing computational power. Researchers are beginning to reassess the role of noise in quantum systems, moving beyond simply trying to eliminate it. Carefully applied noise, specifically amplitude damping, can expand the computational possibilities within a quantum system, accelerating the approach to a fully randomised state. Acknowledging that deliberately adding noise to a quantum system seems counterintuitive given years of effort to minimise it, this work offers an important shift in perspective. Specific quantum noise, namely amplitude damping, alongside standard computational gates can accelerate the creation of highly randomised quantum states. Mathematically demonstrating an effective expansion of the ‘manifold of pure states’, the total set of possible quantum conditions, confirms that noise isn’t always detrimental. This geometric mechanism explains how noise, under certain conditions, enriches the dynamics of quantum systems, moving them faster towards a desired randomised state than previously thought possible. Introducing controlled noise can accelerate quantum computation by expanding the system’s operational space, challenging conventional thinking which prioritised noise reduction as vital for building stable quantum systems.

The research demonstrated that specific noise, in the form of amplitude damping, can accelerate the approach to Haar-like distributions of quantum states. This finding challenges the conventional view of noise as solely detrimental to quantum computation, instead suggesting it can be a resource for enriching system dynamics. By inducing an effective volume expansion on the manifold of pure states, noise allows for quicker exploration of the quantum state space using a single qubit model. The authors propose this geometric mechanism may be beneficial in the design of future quantum algorithms.