Neuromorphic Computing and Spiking Neural Networks for Embodied Intelligence

Embodied intelligence, the capacity of an agent to perceive, learn, and act robustly within a physical environment, remains a central challenge at the intersection of artificial intelligence, robotics, and neuroscience. Conventional deep learning architectures, while powerful, are often computationally expensive, data hungry, and poorly suited to the low latency, low power, and adaptive demands of real world embodied systems. Neuromorphic computing offers a compelling alternative. By emulating the brain’s event driven, sparse, and asynchronous style of computation through spiking neural networks (SNNs), it enables agents to process sensory information and generate motor responses with the efficiency, robustness, and adaptability that characterize biological organisms operating under uncertainty and time pressure.This Research Topic focuses on how neuromorphic computing and spiking neural networks can advance embodied intelligence, that is, intelligent behavior grounded in real time sensorimotor interaction with a dynamic environment. We are particularly interested in work that closes the loop between neuromorphic perception, decision making, and physical action, whether in robotic platforms, autonomous agents, or other embodied systems. Key questions include how spiking neural networks can support robust and efficient sensorimotor control under uncertainty and limited supervision, what architectures and learning rules allow neuromorphic systems to adapt online to novel or changing environments, and how event based sensing can be integrated with spike based decision making to enable fast, energy efficient, and resilient embodied behavior. Addressing these questions requires progress that spans algorithm design, hardware implementation, and system level integration, bringing together researchers from neuromorphic engineering, robotics, and computational neuroscience.We welcome contributions covering, but not limited to, the following themes: - Spiking neural networks for real time sensorimotor control and robotic decision making; - Neuromorphic architectures for embodied perception action loops; event based sensing integrated with spike based planning and control; - On chip and bio plausible learning rules for adaptive embodied agents; - Energy efficient neuromorphic hardware for autonomous robotic platforms; - Multimodal event driven sensory fusion for embodied navigation and interaction; - Benchmarks and evaluation frameworks for neuromorphic embodied intelligence. Through this collection, we aim to advance the design of embodied systems that are efficient, adaptive, and resilient in real world settings. Topic editors: Fan Zhang, Qing Liu, Wenwen Qiang, Junnan Li Source: https://www.frontiersin.org/research-topics/85150/neuromorphic-computing-and-spiking-neural-networks-for-embodied-intelligence