Focus Issue on Algorithms for Neuromorphic Computing

Neuromorphic computing provides a promising energy-efficient alternative to von-Neumann-type computing and learning architectures. However, the best neuromorphic hardware is useless without suitable inference and learning algorithms that can fully exploit hardware advantages. Such algorithms often have to deal with challenging constraints posed by neuromorphic hardware such as massive parallelism, sparse asynchronous communication, and analog and/or unreliable computing elements. This Focus Issue presents advances on various aspects of algorithms for neuromorphic computing. The collection of articles covers a wide range from very fundamental questions about the computational properties of the basic computing elements in neuromorphic systems, algorithms for continual learning, semantic segmentation, and novel efficient learning paradigms, up to algorithms for a specific application domain.

Mixed-signal and fully digital neuromorphic systems have been of significant interest for deploying spiking neural networks in an energy-efficient manner. However, many of these systems impose constraints in terms of fan-in, memory, or synaptic weight precision that have to be considered during network design and training. In this paper, we present quantized rewiring (Q-rewiring), an algorithm that can train both spiking and non-spiking neural networks while meeting hardware constraints during the entire training process. To demonstrate our approach, we train both feedforward and recurrent neural networks with a combined fan-in/weight precision limit, a constraint that is, for example, present in the DYNAP-SE mixed-signal analog/digital neuromorphic processor. Q-rewiring simultaneously performs quantization and rewiring of synapses and synaptic weights through gradient descent updates and projecting the trainable parameters to a constraint-compliant region. Using our algorithm, we find trade-offs between the number of incoming connections to neurons and network performance for a number of common benchmark datasets.

Spiking neural networks (SNNs) are gaining attention due to their energy-efficient computing ability, making them relevant for implementation on low-power neuromorphic hardware. Their biological plausibility has permitted them to benefit from unsupervised learning with bio-inspired plasticity rules, such as spike timing-dependent plasticity (STDP). However, standard STDP has some limitations that make it challenging to implement on hardware. In this paper, we propose a convolutional SNN (CSNN) integrating single-spike integrate-and-fire (SSIF) neurons and trained for the first time with voltage-dependent synaptic plasticity (VDSP), a novel unsupervised and local plasticity rule developed for the implementation of STDP on memristive-based neuromorphic hardware. We evaluated the CSNN on the TIDIGITS dataset, where, helped by our sound preprocessing pipeline, we obtained a performance better than the state of the art, with a mean accuracy of 99.43%. Moreover, the use of SSIF neurons, coupled with time-to-first-spike (TTFS) encoding, results in a sparsely activated model, as we recorded a mean of 5036 spikes per input over the 172 580 neurons of the network. This makes the proposed CSNN promising for the development of models that are extremely efficient in energy. We also demonstrate the efficiency of VDSP on the MNIST dataset, where we obtained results comparable to the state of the art, with an accuracy of 98.56%. Our adaptation of VDSP for SSIF neurons introduces a depression factor that has been very effective at reducing the number of training samples needed, and hence, training time, by a factor of two and more, with similar performance.

Mixed-signal neuromorphic computers often emulate some variant of the LIF neuron model. While, in theory, two-layer networks of these neurons are universal function approximators, single-layer networks consisting of slightly more complex neurons can, at the cost of universality, be more efficient. In this paper, we discuss a family of LIF neurons with passive dendrites. We provide rules that describe how input channels targeting different dendritic compartments interact, and test in how far these interactions can be harnessed in a spiking neural network context. We find that a single layer of two-compartment neurons approximates some functions at smaller errors than similarly sized hidden-layer networks. Single-layer networks with with three compartment neurons can approximate functions such as XOR and four-quadrant multiplication well; adding more compartments only offers small improvements in accuracy. From the perspective of mixed-signal neuromorphic systems, our results suggest that only small modifications to the neuron circuit are necessary to construct more computationally powerful and energy efficient systems that move more computation into the dendritic, analogue domain.

Emerging usages for millimeter wave (mmWave) radar have drawn extensive attention and inspired the exploration of learning mmWave radar data. To be effective, instead of using conventional approaches, recent works have employed modern neural network models to process mmWave radar data. However, due to some inevitable obstacles, e.g., noise and sparsity issues in data, the existing approaches are generally customized for specific scenarios. In this paper, we propose a general neuromorphic framework, termed mm-SNN, to process mmWave radar data with spiking neural networks (SNNs), leveraging the intrinsic advantages of SNNs in processing noisy and sparse data. Specifically, we first present the overall design of mm-SNN, which is adaptive and easily expanded for multi-sensor systems. Second, we introduce general and straightforward attention-based improvements into the mm-SNN to enhance the data representation, helping promote performance. Moreover, we conduct explorative experiments to certify the robustness and effectiveness of the mm-SNN. To the best of our knowledge, mm-SNN is the first SNN-based framework that processes mmWave radar data without using extra modules to alleviate the noise and sparsity issues, and at the same time, achieve considerable performance in the task of trajectory estimation.

There is an increasing need for continual learning in dynamic systems at the edge, such as self-driving vehicles, surveillance drones, and robotic systems. Such a system requires learning from the data stream, training the model to preserve previous information and adapt to a new task, and generating a single-headed vector for future inference, within a limited power budget. Different from previous continual learning algorithms with dynamic structures, this work focuses on a single network and model segmentation to mitigate catastrophic forgetting problem. Leveraging the redundant capacity of a single network, model parameters for each task are separated into two groups: one important group which is frozen to preserve current knowledge, and a secondary group to be saved (not pruned) for future learning. A fixed-size memory containing a small amount of previously seen data is further adopted to assist the training. Without additional regularization, the simple yet effective approach of progressive segmented training (PST) successfully incorporates multiple tasks and achieves state-of-the-art accuracy in the single-head evaluation on the CIFAR-10 and CIFAR-100 datasets. Moreover, the segmented training significantly improves computation efficiency in continual learning and thus, enabling efficient continual learning at the edge. On Intel Stratix-10 MX FPGA, we further demonstrate the efficiency of PST with representative CNNs trained on CIFAR-10.

Spiking neural networks (SNNs) have recently emerged as the low-power alternative to artificial neural networks (ANNs) because of their sparse, asynchronous, and binary event-driven processing. Due to their energy efficiency, SNNs have a high possibility of being deployed for real-world, resource-constrained systems such as autonomous vehicles and drones. However, owing to their non-differentiable and complex neuronal dynamics, most previous SNN optimization methods have been limited to image recognition. In this paper, we explore the SNN applications beyond classification and present semantic segmentation networks configured with spiking neurons. Specifically, we first investigate two representative SNN optimization techniques for recognition tasks (i.e., ANN-SNN conversion and surrogate gradient learning) on semantic segmentation datasets. We observe that, when converted from ANNs, SNNs suffer from high latency and low performance due to the spatial variance of features. Therefore, we directly train networks with surrogate gradient learning, resulting in lower latency and higher performance than ANN-SNN conversion. Moreover, we redesign two fundamental ANN segmentation architectures (i.e., Fully Convolutional Networks and DeepLab) for the SNN domain. We conduct experiments on three semantic segmentation benchmarks including PASCAL VOC2012 dataset, DDD17 event-based dataset, and synthetic segmentation dataset combined CIFAR10 and MNIST datasets. In addition to showing the feasibility of SNNs for semantic segmentation, we show that SNNs can be more robust and energy-efficient compared to their ANN counterparts in this domain.