npj Unconventional Computing最新文献

Neuromorphic (brain-inspired) photonics accelerates AI¹ with high-speed, energy-efficient solutions for RF communication², image processing^3,4, and fast matrix multiplication^5,6. However, integrated neuromorphic photonic hardware faces size constraints that limit network complexity. Recent advances in photonic quantum hardware⁷ and performant trainable quantum circuits⁸ offer a path to more scalable photonic neural networks. Here, we show that a combination of classical network layers with trainable continuous variable quantum circuits yields hybrid networks with improved trainability and accuracy. On a classification task, these hybrid networks match the performance of classical networks nearly twice their size. These performance benefits remain even when evaluated at state-of-the-art bit precisions for classical and quantum hardware. Finally, we outline available hardware and a roadmap to hybrid architectures. These hybrid quantum-classical networks demonstrate a unique route to enhance the computational capacity of integrated photonic neural networks without increasing the network size.

Neuromorphic computing takes biological inspiration to the device level aiming to improve computational efficiency and capabilities. One of the major issues that arises is the training of neuromorphic hardware systems. Typically training algorithms require global information and are thus inefficient to implement directly in hardware. In this paper we describe a set of reinforcement learning based, local weight update rules and their implementation in superconducting hardware. Using SPICE circuit simulations, we implement a small-scale neural network with a learning time of order one nanosecond per update. This network can be trained to learn new functions simply by changing the target output for a given set of inputs, without the need for any external adjustments to the network. Further, this architecture does not require programing explicit weight values in the network, alleviating a critical challenge with analog hardware implementations of neural networks.

With the advent of novel sensor and machine learning technologies, it is becoming possible to develop wearable systems that perform continuous recording and processing of biosignals for health or body state assessment. For example, modern smartwatches can already track physiological functions, including heart rate and its anomalies, with high precision. However, stringent constraints on size and energy consumption pose significant challenges for always-on operation to detect trends across multiple time scales for extended periods of time. To address these challenges, we propose an alternative solution that exploits the ultra-low power consumption features of mixed-signal neuromorphic technologies. We present a biosignal processing architecture that integrates multimodal sensory inputs and processes them using the principles of neural computation to reliably detect trends in heart rate and physiological states. We validate this architecture on a mixed-signal neuromorphic processor and demonstrate its robust operation despite the inherent variability of the analog circuits present in the system. In addition, we demonstrate how the system can process multi scale signals, namely instantaneous heart rate and its long-term states discretized into distinct zones, effectively detecting monotonic changes over extended periods that indicate pathological conditions such as agitation. This approach paves the way for a new generation of energy-efficient stand-alone wearable devices that are particularly suited for scenarios that require continuous health monitoring with minimal device maintenance.