Spike Based
Spike-based neural networks (SNNs) aim to mimic the brain's energy-efficient, event-driven computation by representing information as sequences of spikes. Current research focuses on improving SNN performance through novel architectures like spike transformers and enhanced leaky integrate-and-fire neurons, as well as developing efficient training methods such as those based on implicit differentiation and evolutionary algorithms. This field is significant due to SNNs' potential for low-power applications in brain-computer interfaces, neuromorphic computing, and other areas requiring energy-efficient AI.
Papers
Efficient Speech Command Recognition Leveraging Spiking Neural Network and Curriculum Learning-based Knowledge Distillation
Jiaqi Wang, Liutao Yu, Liwei Huang, Chenlin Zhou, Han Zhang, Zhenxi Song, Min Zhang, Zhengyu Ma, Zhiguo Zhang
CREST: An Efficient Conjointly-trained Spike-driven Framework for Event-based Object Detection Exploiting Spatiotemporal Dynamics
Ruixin Mao, Aoyu Shen, Lin Tang, Jun Zhou