Machine Learning has achieved immense progress by exploiting CPUs and GPUs on classical computing hardware. However, the inevitable end of Moore's Law on these devices requires the adaptation and exploration of novel computational platforms in order to continue these advancements. Biologically accurate, energy efficient neuromorphic systems and fully en-tangled quantum systems are particularly promising arenas for enabling future advances. In this work, we perform a detailed comparison on a level playing field between these two novel substrates by applying them to an identical challenge.We solve the sparse coding problem using the biologically inspired Locally Competitive Algorithm (LCA) on the D-Wave quantum annealer and Intel Loihi neuromorphic spiking processor. The Fashion-MNIST data set was chosen and dimensionally-reduced by sparse Principal Component Analysis (sPCA). A sign flipped second data set was created and appended to the original in order to give each class a mean zero distribution, effectively creating an environment where the data could not be linearly separated. An early in time normalization technique for Loihi is presented along with analysis of optimal parameter selection and unsupervised dictionary learning for all three variations. Studies are ongoing, but preliminary results suggest each computational substrate requires casting the NP-Hard optimization problem in a slightly different manner to best capture the individual strengths, and the new Loihi method allows for more realistic comparison between the two.
Machine Learning has achieved immense progress by exploiting CPUs and GPUs on classical computing hardware. However, the inevitable end of Moore's Law on these devices requires the adaptation and exploration of novel computational platforms in order to continue these advancements. Biologically accurate, energy efficient neuromorphic systems and fully en-tangled quantum systems are particularly promising arenas for enabling future advances. In this work, we perform a detailed comparison on a level playing field between these two novel substrates by applying them to an identical challenge.We solve the sparse coding problem using the biologically inspired Locally Competitive Algorithm (LCA) on the D-Wave quantum annealer and Intel Loihi neuromorphic spiking processor. The Fashion-MNIST data set was chosen and dimensionally-reduced by sparse Principal Component Analysis (sPCA). A sign flipped second data set was created and appended to the original in order to give each class a mean zero distribution, effectively creating an environment where the data could not be linearly separated. An early in time normalization technique for Loihi is presented along with analysis of optimal parameter selection and unsupervised dictionary learning for all three variations. Studies are ongoing, but preliminary results suggest each computational substrate requires casting the NP-Hard optimization problem in a slightly different manner to best capture the individual strengths, and the new Loihi method allows for more realistic comparison between the two.
Although machine learning currently relies on conventional computer architectures, the looming end of Moore's Law necessitates exploration of novel computational platforms. Neuromorphic and quantum systems are a natural path to pursue; biological neurons are incredibly efficient, and quantum mechanics provides theoretical foundations for fast solutions to optimization problems. Here, we make the first comparison of emerging hardware (D-Wave quantum annealer and Intel Loihi spiking processor) on an identically-posed machine learning problem. We implement the bioinspired Locally Competitive Algorithm (LCA) for solving sparse coding on the different substrates. To make the comparison valid, our dataset of choice (Fashion MNIST) is dimensionally-reduced via sparse principal component analysis, under the constraint that both classification performance and a graph-based clustering metric remain unchanged. This enables the problem to be mapped identically to both devices. An analysis of several metrics, including power consumption, reconstruction, and classification accuracy are presented. When given the same specifically-constructed challenge, both substrates perform similarly. Our results suggest while neuromorphic and quantum systems are still in their infancy, they present a possible route to address certain types of classically challenging problems, such as sparse coding, in a way that leverages the unique aspects of the substrates.
Abstract We consider the problem of computing a sparse binary representation of an image. Given an image and an overcomplete, non-orthonormal basis, we aim to find a sparse binary vector indicating the minimal set of basis vectors that when added together best reconstruct the given input. We formulate this problem with an L 2 loss on the reconstruction error, and an L 0 loss on the binary vector enforcing sparsity. First, we solve the sparse representation QUBOs by solving them both on a D-Wave quantum annealer with Pegasus chip connectivity, as well as on the Intel Loihi 2 spiking neuromorphic processor using a stochastic Non-equilibrium Boltzmann Machine (NEBM). Second, using Quantum Evolution Monte Carlo with Reverse Annealing and iterated warm starting on Loihi 2 to evolve the solution quality from the respective machines. We demonstrate that both quantum annealing and neuromorphic computing are suitable for solving binary sparse coding QUBOs.
The exponential growth of large language models (LLMs) and deep neural networks has pushed conventional GPUbased compute infrastructures toward fundamental physical and economic limits. While graphics processing units (GPUs) remainthe dominant platform for AI training and inference, emerging hardware paradigms—particularly quantum computing processors and neuromorphic chips—promise transformative gains in energy efficiency, parallelism, and computational throughput.This paper presents a systematic comparative analysis of three leading hardware classes: high-performance GPUs (NVIDIA H100, AMD MI300X), gate-based quantum processors (IBM Eagle, Google Sycamore), and neuromorphic processors (Intel Loihi 2, IBM TrueNorth). We evaluate these architectures across five critical dimensions: raw throughput, energy efficiency, scalability, programmability, and readiness for production AI workloads. Our analysis demonstrates that while GPUs maintain decisive advantages for transformer-based workloads through 2027, quantum annealers exhibit compelling potential for combinatorial optimization in neural architecture search, and neuromorphic chips offer orders-of-magnitude improvements in energy efficiency for sparse, event-driven inference tasks. We conclude with a proposed hybrid hardware stack and a technology readiness roadmap for the 2025–2035 decade, offering practical guidance for AI systems architects and researchers navigating the post-Moore’s Law compute landscape.
There has been a rapid advancement in developing computational models originating from the governing principles of quantum mechanics. Quantum computing approaches such as variational quantum algorithms (VQAs) and quantum-inspired algorithms have shown advantages across various applications. VQAs have been widely used as hybrid classical-quantum computational frameworks that are well-suited for NISQ advantages. Alternatively, the quantum-inspired evolutionary algorithm (QiEA) provides an efficient search space and parameter optimization solution, useful in handling problems that consist of large parameters. Following the recent trends, quantum-inspired technique, developments, limitations, and future scope are highlighted. Also, an approach towards advancing neuromorphic computation is discussed. Neuromorphic computation and systems are a fast developing domain inherited from fast and massively parallel processing, low energy consumption, high density of millions of neurons in a chip, and have found successful applications across spatio-temporal domain areas including brain data modelling. However, they require a fast and efficient parameter optimization process in large parameter space. Finally, future directions for the development of quantum-enhanced spiking neural networks and neuromorphic systems are outlined.
We use a quantum annealing D-Wave 2X computer to obtain solutions to NP-hard sparse coding problems. To reduce the dimensionality of the sparse coding problem to fit on the quantum D-Wave 2X hardware, we passed downsampled MNIST images through a bottleneck autoencoder. To establish a benchmark for classification performance on this reduced dimensional data set, we built two deep convolutional neural networks (DCNNs). The first DCNN used an AlexNet-like architecture and the second a state-of-the-art residual network (RESNET)model, both implemented in TensorFlow. The two DCNNs yielded classification scores of 94.54 ± 0.7% and 98.8 ± 0.1%, respectively. As a control, we showed that both DCNN architectures produced near-state-of-the-art classification performance ( ~ 99%) on the original MNIST images. To obtain a set of optimized features for inferring sparse representations of the reduced dimensional MNIST dataset, we imprinted on a random set of 47 image patches followed by an off-line unsupervised learning algorithm using stochastic gradient descent to optimize for sparse coding. Our single-layer of sparse coding matched the stride and patch size of the first convolutional layer of the AlexNet-like DCNN and contained 47 fully-connected features, 47 being the maximum number of dictionary elements that could be embedded onto the D-Wave 2X hardware. When the sparse representations inferred by the D-Wave 2X were passed to a linear support vector machine, we obtained a classification score of 95.68%. We found that the classification performance supported by quantum inference was maximal at an optimal level of sparsity corresponding to a critical value of the sparsity/reconstruction error trade-off parameter that previous work has associated with a second order phase transition, an observation supported by a free energy analysis of D-Wave energy states. We mimicked a transfer learning protocol by feeding the D-Wave representations into a multilayer perceptron (MLP), yielding 98.48% classification performance. The classification performance supported by a single-layer of quantum inference was superior to that supported by a classical matching pursuit algorithm set to the same level of sparsity. Whereas the classification performance of both DCNNs declined as the number of training examples was reduced, the classification performance supported by quantum inference was insensitive to the number of training examples. We thus conclude that quantum inference supports classification of reduced dimensional MNIST images exceeding that of a size-matched AlexNet-like DCNN and nearly equivalent to a state-of-the-art RESNET DCNN.
The convergence of quantum computing and neuromorphic engineering represents a promising frontier for next-generation artificial intelligence. Hybrid quantum-neuromorphic intelligence is emerging as an alternative to conventional von Neumann computing by combining the strengths of both paradigms. Quantum systems provide powerful capabilities for optimization, simulation, and complex problem solving, while neuromorphic systems enable energy-efficient, event-driven processing inspired by biological neural networks. Their integration offers new opportunities to address challenges related to computational complexity, scalability, energy consumption, and adaptive learning. Quantum machine learning, quantum neural networks, spiking neural networks, hybrid optimization approaches, and applications in healthcare, smart cities, cybersecurity, and autonomous systems. It also highlights explainability, ethical considerations, and future research directions, concluding that quantum-neuromorphic intelligence could become a key enabler of future AI industries and advanced intelligent systems.