当前位置: X-MOL 学术J. Sign. Process. Syst. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Efficient Neuromorphic Signal Processing with Resonator Neurons
Journal of Signal Processing Systems ( IF 1.6 ) Pub Date : 2022-05-26 , DOI: 10.1007/s11265-022-01772-5
E. Paxon Frady , Sophia Sanborn , Sumit Bam Shrestha , Daniel Ben Dayan Rubin , Garrick Orchard , Friedrich T. Sommer , Mike Davies

The biologically inspired spiking neurons used in neuromorphic computing are nonlinear filters with dynamic state variables, which is distinct from the stateless neuron models used in deep learning. The new version of Intel’s neuromorphic research processor, Loihi 2, supports an extended range of stateful spiking neuron models with programmable dynamics. Here, we showcase advanced neuron models that can be used to efficiently process streaming data in simulation experiments on emulated Loihi 2 hardware. In one example, Resonate-and-Fire (RF) neurons are used to compute the Short Time Fourier Transform (STFT) with similar computational complexity but 47x less output bandwidth than the conventional STFT. In another example, we describe an algorithm for optical flow estimation using spatiotemporal RF neurons that requires over 90x fewer operations than a conventional DNN-based solution. We also demonstrate backpropagation methods to train non-linear spiking RF neurons for audio classification tasks, suitable for efficient execution on Loihi 2. We conclude with another application of nonlinear filtering showing a cascade of Hopf resonators exhibiting computational properties seen in the cochlea, such as self-normalization. Taken together, this work presents new techniques for an efficient spike-based spectrogram encoder that can be used for signal processing applications.



中文翻译:

使用谐振神经元进行高效的神经形态信号处理

神经形态计算中使用的受生物学启发的尖峰神经元是具有动态状态变量的非线性滤波器,这与深度学习中使用的无状态神经元模型不同。英特尔神经形态研究处理器的新版本 Loihi 2 支持范围更广的具有可编程动态的状态脉冲神经元模型。在这里,我们展示了可用于在仿真 Loihi 2 硬件上的模拟实验中有效处理流数据的高级神经元模型。在一个示例中,共振激发 (RF) 神经元用于计算短时傅里叶变换 (STFT),其计算复杂度相似,但输出带宽比传统 STFT 少 47 倍。在另一个例子中,我们描述了一种使用时空 RF 神经元进行光流估计的算法,该算法所需的操作比传统的基于 DNN 的解决方案少 90 倍以上。我们还展示了反向传播方法来训练非线性尖峰射频神经元用于音频分类任务,适用于 Loihi 2 上的高效执行。我们以非线性滤波的另一个应用结束,展示了在耳蜗中看到的表现出计算特性的级联 Hopf 谐振器,例如自归一化。总而言之,这项工作提出了一种可用于信号处理应用的高效基于尖峰的频谱图编码器的新技术。我们以非线性滤波的另一个应用结束,该应用显示了一系列 Hopf 谐振器,这些谐振器表现出耳蜗中的计算特性,例如自归一化。总而言之,这项工作提出了一种可用于信号处理应用的高效基于尖峰的频谱图编码器的新技术。我们以非线性滤波的另一个应用结束,该应用显示了一系列 Hopf 谐振器,这些谐振器表现出耳蜗中的计算特性,例如自归一化。总而言之,这项工作提出了一种可用于信号处理应用的高效基于尖峰的频谱图编码器的新技术。

更新日期:2022-05-27
down
wechat
bug