The deployment process of a spiking neural network (SNN) often involves partitioning the neural network and mapping these partitions onto processing units within the neuromorphic hardware. Finding optimal deployment schemes is an NP-hard problem. Optimizing these schemes presents challenges, particular in devising computationally effective cost functions optimization objectives such as communication time consumption and energy efficiency. These objectives require consideration of network dynamics shaped by neuron activity patterns, demanding intricate mathematical analyses or simulations for integrating them into a cost model for SNN development. Our approach focuses on network dynamics, which are hardware-independent and can be modeled separately from specific hardware configurations. We employ a pairwise Ising-type maximum entropy model, which is a model show effective in accurately capturing pairwise correlations among system components in a collaborative system. On top of this model, we incorporates hardware and network structure-specific factors to devise a cost function. We conducted an extremely preliminary investigation using the SpiNNaker machine. We show that the ising model training can also be computationally complex. Currently, we lack sufficient evidence to substantiate the effectiveness of our proposed methods. Further efforts is needed to explore integrating network dynamics into SNN deployment.

翻译：脉冲神经网络（SNN）的部署过程通常涉及对神经网络进行分区，并将这些分区映射到神经形态硬件内的处理单元上。寻找最优部署方案是一个NP难问题。优化这些方案面临挑战，特别是在设计计算有效的成本函数优化目标时，例如通信时间消耗和能源效率。这些目标需要考虑由神经元活动模式塑造的网络动态，这需要复杂的数学分析或模拟才能将其整合到SNN部署的成本模型中。我们的方法侧重于网络动态，这些动态与硬件无关，可以独立于特定的硬件配置进行建模。我们采用了一种成对伊辛型最大熵模型，该模型在协作系统中被证明能有效捕捉系统组件间的成对相关性。在此模型基础上，我们结合了硬件和网络结构特定的因素来设计成本函数。我们使用SpiNNaker机器进行了极其初步的探索。我们发现伊辛模型的训练在计算上也可能很复杂。目前，我们缺乏足够的证据来证实所提方法的有效性。需要进一步努力探索将网络动态整合到SNN部署中。