Spiking neural networks (SNNs) are a promising avenue for machine learning with superior energy efficiency compared to traditional artificial neural networks (ANNs). Recent advances in training and ...input encoding have put SNNs on par with state-of-the-art ANNs in image classification. However, such tasks do not utilize the internal dynamics of SNNs fully. Notably, a spiking neuron's membrane potential acts as an internal memory, merging incoming inputs sequentially. This recurrent dynamic enables the networks to learn temporal correlations, making SNNs suitable for sequential learning. Such problems can also be tackled using ANNs. However, to capture the temporal dependencies, either the inputs have to be lumped over time (e.g. Transformers); or explicit recurrence needs to be introduced (e.g. recurrent neural networks (RNNs), long-short-term memory (LSTM) networks), which incurs considerable complexity. To that end, we explore the capabilities of SNNs in providing lightweight solutions to four sequential tasks involving text, speech and vision. Our results demonstrate that SNNs, by leveraging their intrinsic memory, can be an efficient alternative to RNNs and LSTMs for sequence processing, especially for certain edge applications. Furthermore, SNNs can be combined with ANNs (hybrid networks) synergistically to obtain the best of both worlds in terms of accuracy and efficiency.
Surgical action localization is a challenging computer vision problem. While it has promising applications including automated training of surgery procedures, surgical workflow optimization, etc., ...appropriate model design is pivotal to accomplishing this task. Moreover, the lack of suitable medical datasets adds an additional layer of complexity. To that effect, we introduce a new complex dataset of nephrectomy surgeries called UroSlice. To perform the action localization from these videos, we propose a novel model termed as `ViTALS' (Vision Transformer for Action Localization in Surgical Nephrectomy). Our model incorporates hierarchical dilated temporal convolution layers and inter-layer residual connections to capture the temporal correlations at finer as well as coarser granularities. The proposed approach achieves state-of-the-art performance on Cholec80 and UroSlice datasets (89.8% and 66.1% accuracy, respectively), validating its effectiveness.
The particle swarm optimizer (PSO) is a swarm intelligence based on heuristic optimization technique that can be applied to a wide range of problems. After analyzing the dynamics of traditional PSO, ...this paper presents a new PSO variant on the basis of local stochastic search strategy (LSSPSO) for performance improvement. This is encouraged by a social phenomenon that everyone wants to first exceed the nearest superior and then all superior. Specifically, LSSPSO employs a local stochastic search to adjust inertia weight in terms of keeping a balance between the diversity and the convergence speed, aiming to improve the performance of traditional PSO. Experiments conducted on unimodal and multimodal test functions demonstrate the effectiveness of LSSPSO in solving multiple benchmark problems as compared to several other PSO variants.
Cognitive radio (CR) networks will enable dynamic spectrum re-use and thereby accelerate the adoption of high bandwidth services in available licensed frequencies with better channel characteristics. ...However, the possibility of the licensed user reclaiming the channel raises additional concerns on how best to reserve resources for secondary users (SUs) that are likely to have different qualities of service (QoSs) depending on their application requirements. This paper addresses the problem of spectrum resource management for co-located SUs with both streaming and intermittent data by efficiently identifying the number of backup channels that will ensure seamless end to end service. The contributions of this paper are threefold: First, a comprehensive analytical framework based on queueing theory is devised to calculate the theoretical delay in accessing the spectrum depending on the required QoS, with guidelines on how to optimize the set of back-up channels for possible future use; second, a method of spectrum allocation for SUs with these different QoS demands is formulated, especially as they co-exist and affect the performance of each other; third, a case study of applying these techniques in a novel application area of wireless medical telemetry is presented. Results reveal that the simulated spectral efficiency of the channel allocation using our approach matches closely with our theoretical predictions, within a 5% bound.
Spiking Neural Networks (SNNs) are energy efficient alternatives to commonly used deep neural networks (DNNs). Through event-driven information processing, SNNs can reduce the expensive compute ...requirements of DNNs considerably, while achieving comparable performance. However, high inference latency is a significant hindrance to the edge deployment of deep SNNs. Computation over multiple timesteps not only increases latency as well as overall energy budget due to higher number of operations, but also incurs memory access overhead of fetching membrane potentials, both of which lessen the energy benefits of SNNs. To overcome this bottleneck and leverage the full potential of SNNs, we propose an Iterative Initialization and Retraining method for SNNs (IIR-SNN) to perform single shot inference in the temporal axis. The method starts with an SNN trained with T timesteps (T>1). Then at each stage of latency reduction, the network trained at previous stage with higher timestep is utilized as initialization for subsequent training with lower timestep. This acts as a compression method, as the network is gradually shrunk in the temporal domain. In this paper, we use direct input encoding and choose T=5, since as per literature, it is the minimum required latency to achieve satisfactory performance on ImageNet. The proposed scheme allows us to obtain SNNs with up to unit latency, requiring a single forward pass during inference. We achieve top-1 accuracy of 93.05%, 70.15% and 67.71% on CIFAR-10, CIFAR-100 and ImageNet, respectively using VGG16, with just 1 timestep. In addition, IIR-SNNs perform inference with 5-2500X reduced latency compared to other state-of-the-art SNNs, maintaining comparable or even better accuracy. Furthermore, in comparison with standard DNNs, the proposed IIR-SNNs provide25-33X higher energy efficiency, while being comparable to them in classification performance.
Spiking Neural Networks (SNNs) are a promising alternative to traditional deep learning methods since they perform event-driven information processing. However, a major drawback of SNNs is high ...inference latency. The efficiency of SNNs could be enhanced using compression methods such as pruning and quantization. Notably, SNNs, unlike their non-spiking counterparts, consist of a temporal dimension, the compression of which can lead to latency reduction. In this paper, we propose spatial and temporal pruning of SNNs. First, structured spatial pruning is performed by determining the layer-wise significant dimensions using principal component analysis of the average accumulated membrane potential of the neurons. This step leads to 10-14X model compression. Additionally, it enables inference with lower latency and decreases the spike count per inference. To further reduce latency, temporal pruning is performed by gradually reducing the timesteps while training. The networks are trained using surrogate gradient descent based backpropagation and we validate the results on CIFAR10 and CIFAR100, using VGG architectures. The spatiotemporally pruned SNNs achieve 89.04% and 66.4% accuracy on CIFAR10 and CIFAR100, respectively, while performing inference with 3-30X reduced latency compared to state-of-the-art SNNs. Moreover, they require 8-14X lesser compute energy compared to their unpruned standard deep learning counterparts. The energy numbers are obtained by multiplying the number of operations with energy per operation. These SNNs also provide 1-4% higher robustness against Gaussian noise corrupted inputs. Furthermore, we perform weight quantization and find that performance remains reasonably stable up to 5-bit quantization.
Spiking Neural Networks (SNNs) offer a promising alternative to traditional deep learning frameworks, since they provide higher computational efficiency due to event-driven information processing. ...SNNs distribute the analog values of pixel intensities into binary spikes over time. However, the most widely used input coding schemes, such as Poisson based rate-coding, do not leverage the additional temporal learning capability of SNNs effectively. Moreover, these SNNs suffer from high inference latency which is a major bottleneck to their deployment. To overcome this, we propose a scalable time-based encoding scheme that utilizes the Discrete Cosine Transform (DCT) to reduce the number of timesteps required for inference. DCT decomposes an image into a weighted sum of sinusoidal basis images. At each time step, the Hadamard product of the DCT coefficients and a single frequency base, taken in order, is given to an accumulator that generates spikes upon crossing a threshold. We use the proposed scheme to learn DCT-SNN, a low-latency deep SNN with leaky-integrate-and-fire neurons, trained using surrogate gradient descent based backpropagation. We achieve top-1 accuracy of 89.94%, 68.3% and 52.43% on CIFAR-10, CIFAR-100 and TinyImageNet, respectively using VGG architectures. Notably, DCT-SNN performs inference with 2-14X reduced latency compared to other state-of-the-art SNNs, while achieving comparable accuracy to their standard deep learning counterparts. The dimension of the transform allows us to control the number of timesteps required for inference. Additionally, we can trade-off accuracy with latency in a principled manner by dropping the highest frequency components during inference.
Spiking Neural Networks (SNNs) are being explored to emulate the astounding capabilities of human brain that can learn and compute functions robustly and efficiently with noisy spiking activities. A ...variety of spiking neuron models have been proposed to resemble biological neuronal functionalities. With varying levels of bio-fidelity, these models often contain a leak path in their internal states, called membrane potentials. While the leaky models have been argued as more bioplausible, a comparative analysis between models with and without leak from a purely computational point of view demands attention. In this paper, we investigate the questions regarding the justification of leak and the pros and cons of using leaky behavior. Our experimental results reveal that leaky neuron model provides improved robustness and better generalization compared to models with no leak. However, leak decreases the sparsity of computation contrary to the common notion. Through a frequency domain analysis, we demonstrate the effect of leak in eliminating the high-frequency components from the input, thus enabling SNNs to be more robust against noisy spike-inputs.