The Power of Annealing Quantum Computing
Yesterday, I hosted Dr. Alan Baratz, CEO, and Dr. Trevor Lanting, CDO of D-Wave, at a Forbes Technology Council event, The Power of Quantum Annealing. The event was marked by rich discussion, and here I summarize a few points.
D-Wave is the only quantum computing company that can state that it offers enterprise-scale quantum computers in operational use with proven, quantifiable ROI. This is because annealing quantum technology is easier to work with and to scale and is subject to much less noise.
A bifurcation in the quantum computing industry was initially caused by the belief that gate models will be good at all quantum computing tasks and that annealing will solve optimization problems. This belief drove most players to pursue the gate model approach. As a result, D-Wave is now the primary player in the annealing quantum computing space. This has been further reinforced by the realization that the gate-based approach requires complex algorithms and careful error correction for optimization versus annealing, which solves optimization by finding the lowest energy state, making it infeasible from a cost perspective for gate models to focus on optimization.
Based on insight gained from five incarnations of D-Wave’s annealing quantum computer, their research efforts are now focused on i) increasing connectivity between qubits, which allows representation of increasingly complicated optimization problems, and ii) increasing coherence, which is critical for quantum effects like superposition, entanglement, and tunneling, which are, in turn, crucial in computations.
Further, continuing quadratic speed up in the ability to reach low-energy solutions proves that D-Wave’s systems harness quantum coherence to deliver computational advantage. Earlier this year, a paper was submitted to a scientific journal showing quantum supremacy in a real-world situation — finding low-energy configurations for Ising spin glass problems posed to different lattice sizes. Computations on D-Wave Advantage2 quantum computers demonstrated an exponential advantage over state-of-the-art classical techniques such as tensor networks, neural networks, and heuristics, which, when performed on Frontier and Summit supercomputers at Oak Ridge National Lab, proved too slow and power-hungry for all but the smallest instances.
This kind of quantum computational advantage will prove to be exceedingly crucial in generative AI, as it will help manage the cost of training, inference, and energy.
My take on why supremacy should not be surprising in a D-Wave type annealing quantum computer compared to gate-based quantum computers is that it derives from a complex adaptive systems paradigm. Leveraging complex adaptive systems—where deeper dynamics range from interconnectivity and self-organization to non-linearity and emergence—to frame problems and generate breakthrough solutions cannot but be vastly superior to a linear paradigm.
The Forbes page and one-hour recording of the event is here:
