Spiking neural networks (SNNs) well support spatiotemporal learning and energyefficient eventdriven hardware neuromorphic processors. As an important class of SNNs, recurrent spiking neural networks (RSNNs) possess great computational power. However, the practical application of RSNNs is severely limited by challenges in training. Biologicallyinspired unsupervised learning has limited capability in boosting the performance of RSNNs. On the other hand, existing backpropagation (BP) methods suffer from high complexity of unfolding in time, vanishing and exploding gradients, and approximate differentiation of discontinuous spiking activities when applied to RSNNs. To enable supervised training of RSNNs under a welldefined loss function, we present a novel SpikeTrain level RSNNs Backpropagation (STRSBP) algorithm for training deep RSNNs. The proposed STRSBP directly computes the gradient of a ratecoded loss function defined at the output layer of the network w.r.t tunable parameters. The scalability of STRSBP is achieved by the proposed spiketrain level computation during which temporal effects of the SNN is captured in both the forward and backward pass of BP. Our STRSBP algorithm can be broadly applied to RSNNs with a single recurrent layer or deep RSNNs with multiple feedforward and recurrent layers. Based upon challenging speech and image datasets including TI46, NTIDIGITS, FashionMNIST and MNIST, STRSBPmore »
Approximating Backpropagation for a Biologically Plausible Local Learning Rule in Spiking Neural Networks
Asynchronous eventdriven computation and communication using spikes facilitate the realization of spiking neural networks (SNN) to be massively parallel, extremely energy efficient and highly robust on specialized neuromorphic hardware. However, the lack of a unified robust learning algorithm limits the SNN to shallow networks with low accuracies. Artificial neural networks (ANN), however, have the backpropagation algorithm which can utilize gradient descent to train networks which are locally robust universal function approximators. But backpropagation algorithm is neither biologically plausible nor neuromorphic implementation friendly because it requires: 1) separate backward and forward passes, 2) differentiable neurons, 3) highprecision propagated errors, 4) coherent copy of weight matrices at feedforward weights and the backward pass, and 5) nonlocal weight update. Thus, we propose an approximation of the backpropagation algorithm completely with spiking neurons and extend it to a local weight update rule which resembles a biologically plausible learning rule spiketimingdependent plasticity (STDP). This will enable error propagation through spiking neurons for a more biologically plausible and neuromorphic implementation friendly backpropagation algorithm for SNNs. We test the proposed algorithm on various traditional and nontraditional benchmarks with competitive results.
 Award ID(s):
 1822165
 Publication Date:
 NSFPAR ID:
 10188106
 Journal Name:
 International Conference on Neuromorphic Systems
 Page Range or eLocationID:
 1 to 8
 Sponsoring Org:
 National Science Foundation
More Like this


We use a recently developed synchronous Spiking Neural Network (SNN) model to study the problem of learning hierarchicallystructured concepts. We introduce an abstract data model that describes simple hierarchical concepts. We define a feedforward layered SNN model, with learning modeled using Oja’s local learning rule, a well known biologicallyplausible rule for adjusting synapse weights. We define what it means for such a network to recognize hierarchical concepts; our notion of recognition is robust, in that it tolerates a bounded amount of noise. Then, we present a learning algorithm by which a layered network may learn to recognize hierarchical concepts according to our robust definition. We analyze correctness and performance rigorously; the amount of time required to learn each concept, after learning all of the subconcepts, is approximately O ( 1ηk(`max log(k) + 1ε) + b log(k)), where k is the number of subconcepts per concept, `max is the maximum hierarchical depth, η is the learning rate, ε describes the amount of uncertainty allowed in robust recognition, and b describes the amount of weight decrease for "irrelevant" edges. An interesting feature of this algorithm is that it allows the network to learn subconcepts in a highly interleaved manner. This algorithm assumesmore »

We use a recently developed synchronous Spiking Neural Network (SNN) model to study the problem of learning hierarchicallystructured concepts. We introduce an abstract data model that describes simple hierarchical concepts. We define a feedforward layered SNN model, with learning modeled using Oja’s local learning rule, a well known biologicallyplausible rule for adjusting synapse weights. We define what it means for such a network to recognize hierarchical concepts; our notion of recognition is robust, in that it tolerates a bounded amount of noise. Then, we present a learning algorithm by which a layered network may learn to recognize hierarchical concepts according to our robust definition. We an alyze correctness and performance rigorously; the amount of time required to learn each concept, after learning all of the subconcepts, is approximately O( 1ηk(`max log(k) + 1ε) + b log(k)), where k is the number of subconcepts per concept, `max is the maximum hierarchical depth, η is the learning rate, ε describes the amount of uncertainty allowed in robust recognition, and b describes the amount of weight decrease for "irrelevant" edges. An interesting feature of this algorithm is that it allows the network to learn subconcepts in a highly interleaved manner. This algorithm assumesmore »

The neural plausibility of backpropagation has long been disputed, primarily for its use of nonlocal weight transport — the biologically dubious requirement that one neuron instantaneously measure the synaptic weights of another. Until recently, attempts to create local learning rules that avoid weight transport have typically failed in the largescale learning scenarios where backpropagation shines, e.g. ImageNet categorization with deep convolutional networks. Here, we investigate a recently proposed local learning rule that yields competitive performance with backpropagation and find that it is highly sensitive to metaparameter choices, requiring laborious tuning that does not transfer across network architecture. Our analysis indicates the underlying mathematical reason for this instability, allowing us to identify a more robust local learning rule that better transfers without metaparameter tuning. Nonetheless, we find a performance and stability gap between this local rule and backpropagation that widens with increasing model depth. We then investigate several nonlocal learning rules that relax the need for instantaneous weight transport into a more biologicallyplausible "weight estimation" process, showing that these rules match stateoftheart performance on deep networks and operate effectively in the presence of noisy updates. Taken together, our results suggest two routes towards the discovery of neural implementations for credit assignmentmore »

A new class of neuromorphic processors promises to provide fast and powerefficient execution of spiking neural networks with onchip synaptic plasticity. This efficiency derives in part from the finegrained parallelism as well as eventdriven communication mediated by spatially and temporally sparse spike messages. Another source of efficiency arises from the close spatial proximity between synapses and the sites where their weights are applied and updated. This proximity of compute and memory elements drastically reduces expensive data movements but imposes the constraint that only local operations can be efficiently performed, similar to constraints present in biological neural circuits. Efficient weight update operations should therefore only depend on information available locally at each synapse as nonlocal operations that involve copying, taking a transpose, or normalizing an entire weight matrix are not efficiently supported by present neuromorphic architectures. Moreover, spikes are typically nonnegative events, which imposes additional constraints on how local weight update operations can be performed. The Locally Competitive Algorithm (LCA) is a dynamical sparse solver that uses only local computations between nonspiking leaky integrator neurons, allowing for massively parallel implementations on compatible neuromorphic architectures such as Intel's Loihi research chip. It has been previously demonstrated that nonspiking LCA can be usedmore »