Editorial: Algorithm-hardware co-optimization in neuromorphic computing for efficient AI

Neuromorphic computing holds the promise of sustainable AI by combining brain-inspired models 4 with event-driven, massively parallel hardware. However, a central question remains: when and how 5 do neuromorphic systems convert their architectural advantages into effective end-to-end efficiency 6 for real-world tasks? This Research Topic presents six contributions that address this question from 7 various perspectives. Specifically, it explores training methods that minimize timesteps and memory 8 usage, hardware-aware algorithms and quantization, emulation techniques that mitigate risks associated 9 with analog platforms, and mapping and scheduling strategies that enhance utilization on many-core 10 neuromorphic chips. 11We are pleased to present a series of innovative research articles in this field that introduce the following