The field of neuromorphic computing is moving towards the development of more efficient and specialized hardware and software solutions. Researchers are focusing on creating novel training algorithms and models that can take advantage of the unique properties of neuromorphic devices, such as spiking neurons and synapses. This includes the development of new models and frameworks for understanding and optimizing the behavior of these devices. Noteworthy papers include:

• Bruno: Backpropagation Running Undersampled for Novel device Optimization, which presents a new approach to training neural networks for hardware based on spiking neurons and synapses.
• Time to Spike? Understanding the Representational Power of Spiking Neural Networks in Discrete Time, which provides a theoretical analysis of the representational power of discrete-time spiking neural networks.