Constructing energy-efficient mixed-precision neural networks through principal component analysis for edge intelligence

DIET-SNN: A Low-Latency Spiking Neural Network With Direct Input Encoding and Leakage and Threshold Optimization

Bioinspired spiking neural networks (SNNs), operating with asynchronous binary signals (or spikes) distributed over time, can potentially lead to greater computational efficiency on event-driven hardware. The state-of-the-art SNNs suffer from high inference latency, resulting from inefficient input encoding and suboptimal settings of the neuron parameters (firing threshold and membrane leak). We propose DIET-SNN, a low-latency deep spiking network trained with gradient descent to optimize the membrane leak and the firing threshold along with other network parameters (weights). The membrane leak and threshold of each layer are optimized with end-to-end backpropagation to achieve competitive accuracy at reduced latency. The input layer directly processes the analog pixel values of an image without converting it to spike train. The first convolutional layer converts analog inputs into spikes where leaky-integrate-and-fire (LIF) neurons integrate the weighted inputs and generate an output spike when the membrane potential crosses the trained firing threshold. The trained membrane leak selectively attenuates the membrane potential, which increases activation sparsity in the network. The reduced latency combined with high activation sparsity provides massive improvements in computational efficiency. We evaluate DIET-SNN on image classification tasks from CIFAR and ImageNet datasets on VGG and ResNet architectures. We achieve top-1 accuracy of 69% with five timesteps (inference latency) on the ImageNet dataset with 12× less compute energy than an equivalent standard artificial neural network (ANN). In addition, DIET-SNN performs 20- 500× faster inference compared to other state-of-the-art SNN models.

Analyzing point cloud of coal mining process in much dust environment based on dynamic graph convolution neural network

Environmental perception is an important research direction of coal mine sustainable development. There is much dust in the underground working environment of coal mine. This study is to identify the marker (ball) in the coal mine, which provides a basic to convert the coordinate of large-scale fully mechanized mining face point cloud to the geodetic coordinate. Firstly, in the face of the phenomenon that the uneven distribution of underground point cloud is more serious, this study further has studied on the basis of complete and incomplete geometry point cloud and generated multi-density geometry point cloud for the first time. Secondly, aiming at the problem that the geometric features of underground point cloud are not obvious enough, this study has increased the weight of point cloud normal vector in the training process of network model, so that the network model is more sensitive to different geometric features. Finally, this study has used a variety of advanced deep neural networks to directly analyze point clouds to verify the proposed method. The results show that the method proposed in this study has been combined with the dynamic graph convolution neural network (DGCNN) established earlier, which can more accurately identify the ball in tens of millions of the point clouds of coal mining process. Most importantly, this work is not only of great significance to improve the production efficiency and safety in fully mechanized mining face but also lays a foundation for realizing intelligence in the mining field and avoiding the harm of dust explosion and other accidents to workers.

Lead federated neuromorphic learning for wireless edge artificial intelligence

In order to realize the full potential of wireless edge artificial intelligence (AI), very large and diverse datasets will often be required for energy-demanding model training on resource-constrained edge devices. This paper proposes a lead federated neuromorphic learning (LFNL) technique, which is a decentralized energy-efficient brain-inspired computing method based on spiking neural networks. The proposed technique will enable edge devices to exploit brain-like biophysiological structure to collaboratively train a global model while helping preserve privacy. Experimental results show that, under the situation of uneven dataset distribution among edge devices, LFNL achieves a comparable recognition accuracy to existing edge AI techniques, while substantially reducing data traffic by >3.5× and computational latency by >2.0×. Furthermore, LFNL significantly reduces energy consumption by >4.5× compared to standard federated learning with a slight accuracy loss up to 1.5%. Therefore, the proposed LFNL can facilitate the development of brain-inspired computing and edge AI.

Designing energy-efficient computing solution for the implementation of AI algorithms in edge devices remains a challenge. Yang et al. proposes a decentralized brain-inspired computing method enabling multiple edge devices to collaboratively train a global model without a fixed central coordinator.

Neuroevolution Guided Hybrid Spiking Neural Network Training

Neuromorphic computing algorithms based on Spiking Neural Networks (SNNs) are evolving to be a disruptive technology driving machine learning research. The overarching goal of this work is to develop a structured algorithmic framework for SNN training that optimizes unique SNN-specific properties like neuron spiking threshold using neuroevolution as a feedback strategy. We provide extensive results for this hybrid bio-inspired training strategy and show that such a feedback-based learning approach leads to explainable neuromorphic systems that adapt to the specific underlying application. Our analysis reveals 53.8, 28.8, and 28.2% latency improvement for the neuroevolution-based SNN training strategy on CIFAR-10, CIFAR-100, and ImageNet datasets, respectively in contrast to state-of-the-art conversion based approaches. The proposed algorithm can be easily extended to other application domains like image classification in presence of adversarial attacks where 43.2 and 27.9% latency improvements were observed on CIFAR-10 and CIFAR-100 datasets, respectively.

A Systematic Literature Review on Distributed Machine Learning in Edge Computing

Distributed edge intelligence is a disruptive research area that enables the execution of machine learning and deep learning (ML/DL) algorithms close to where data are generated. Since edge devices are more limited and heterogeneous than typical cloud devices, many hindrances have to be overcome to fully extract the potential benefits of such an approach (such as data-in-motion analytics). In this paper, we investigate the challenges of running ML/DL on edge devices in a distributed way, paying special attention to how techniques are adapted or designed to execute on these restricted devices. The techniques under discussion pervade the processes of caching, training, inference, and offloading on edge devices. We also explore the benefits and drawbacks of these strategies.

Medical Surgical Video Recognition and Retrieval Based on Novel Unified Approximation

Video retrieval recognition is a linear characterized action constituted by many frame similarity-based videos. This medical video recognition and classification can be a great extent in medical research, such as Endoscopic, radiological, pathological, and applied health informatics. General Video Retrieval Recognition (GVRR) cannot address a problem with recognition alone. GVRR can be solving the Multi-Input-Multi-Output (MIMO) interface mixed video retrieval system. To generalize the conventional video retrieval interface like Multi-user MIMO, WiMAX MIMO, single-user MIMO, several types of research made excused. In fine-tuning existing video retrieval, this research gives the authentic procedure for a frame-based cognitive operation called Secure Approximation and sTability Based Secure Video Retrieval recognition (SAT-SR) recognition proposed. In this research article, the process of recognition has three processes generalized by the video retrieval system. Initially, the virtual dissection and connection weights of input video were established using the mathematical and numerical analysis of interpolation estimation. Secondly, the interpolation approximation and activation function were figured out using the Open Mcrypt Stimulus (oMs) for video security fragments. Similarly, systematic investigations are accomplished for approximation error computation. The result for this widely circulated utilization of three processes on the video retrieval recognition prevents the occurrence of the cybercrime abuse of stored video registers. The proposed technique was used to identify the virtual dissection, interpolation, and activation function for decoding the videos. Using this information, the abusers identified cybercrime rate might be reduced considerably.

Exploring the Connection Between Binary and Spiking Neural Networks

On-chip edge intelligence has necessitated the exploration of algorithmic techniques to reduce the compute requirements of current machine learning frameworks. This work aims to bridge the recent algorithmic progress in training Binary Neural Networks and Spiking Neural Networks-both of which are driven by the same motivation and yet synergies between the two have not been fully explored. We show that training Spiking Neural Networks in the extreme quantization regime results in near full precision accuracies on large-scale datasets like CIFAR-100 and ImageNet. An important implication of this work is that Binary Spiking Neural Networks can be enabled by "In-Memory" hardware accelerators catered for Binary Neural Networks without suffering any accuracy degradation due to binarization. We utilize standard training techniques for non-spiking networks to generate our spiking networks by conversion process and also perform an extensive empirical analysis and explore simple design-time and run-time optimization techniques for reducing inference latency of spiking networks (both for binary and full-precision models) by an order of magnitude over prior work. Our implementation source code and trained models are available at https://github.com/NeuroCompLab-psu/SNN-Conversion.