Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userComplexity-aware Adaptive Training and Inference for Edge-Cloud Distributed AI Systems
91
0
0.0
(
0
)
Added by
Yinghan Long
Publication date
2021
fields
Informatics Engineering
and research's language is
English
Ask ChatGPT about the research
No Arabic abstract
The ubiquitous use of IoT and machine learning applications is creating large amounts of data that require accurate and real-time processing. Although edge-based smart data processing can be enabled by deploying pretrained models, the energy and memory constraints of edge devices necessitate distributed deep learning between the edge and the cloud for complex data. In this paper, we propose a distributed AI system to exploit both the edge and the cloud for training and inference. We propose a new architecture, MEANet, with a main block, an extension block, and an adaptive block for the edge. The inference process can terminate at either the main block, the extension block, or the cloud. The MEANet is trained to categorize inputs into easy/hard/complex classes. The main block identifies instances of easy/hard classes and classifies easy classes with high confidence. Only data with high probabilities of belonging to hard classes would be sent to the extension block for prediction. Further, only if the neural network at the edge shows low confidence in the prediction, the instance is considered complex and sent to the cloud for further processing. The training technique lends to the majority of inference on edge devices while going to the cloud only for a small set of complex jobs, as determined by the edge. The performance of the proposed system is evaluated via extensive experiments using modified models of ResNets and MobileNetV2 on CIFAR-100 and ImageNet datasets. The results show that the proposed distributed model has improved accuracy and energy consumption, indicating its capacity to adapt.
rate research
Read More
Neural networks training on edge terminals is essential for edge AI computing, which needs to be adaptive to evolving environment. Quantised models can efficiently run on edge devices, but existing training methods for these compact models are designed to run on powerful servers with abundant memory and energy budget. For example, quantisation-aware training (QAT) method involves two copies of model parameters, which is usually beyond the capacity of on-chip memory in edge devices. Data movement between off-chip and on-chip memory is energy demanding as well. The resource requirements are trivial for powerful servers, but critical for edge devices. To mitigate these issues, We propose Resource Constrained Training (RCT). RCT only keeps a quantised model throughout the training, so that the memory requirements for model parameters in training is reduced. It adjusts per-layer bitwidth dynamically in order to save energy when a model can learn effectively with lower precision. We carry out experiments with representative models and tasks in image application and natural language processing. Experiments show that RCT saves more than 86% energy for General Matrix Multiply (GEMM) and saves more than 46% memory for model parameters, with limited accuracy loss. Comparing with QAT-based method, RCT saves about half of energy on moving model parameters.
Our objective will be to integrate ML into Fermilab accelerator operations and furthermore provide an accessible framework which can also be used by a broad range of other accelerator systems with dynamic tuning needs. We will develop of real-time accelerator control using embedded ML on-chip hardware and fast communication between distributed systems in this proposal. We will demonstrate this technology for the Mu2e experiment by increasing the overall duty factor and uptime of the experiment through two synergistic projects. First, we will use deep reinforcement learning techniques to improve the performance of the regulation loop through guided optimization to provide stable proton beams extracted from the Delivery Ring to the Mu2e experiment. This requires the development of a digital twin of the system to model the accelerator and develop real-time ML algorithms. Second, we will use de-blending techniques to disentangle and classify overlapping beam losses in the Main Injector and Recycler Ring to reduce overall beam downtime in each machine. This ML model will be deployed within a semi-autonomous operational mode. Both applications require processing at the millisecond scale and will share similar ML-in-hardware techniques and beam instrumentation readout technology. A collaboration between Fermilab and Northwestern University will pull together the talents and resources of accelerator physicists, beam instrumentation engineers, embedded system architects, FPGA board design experts, and ML experts to solve complex real-time accelerator controls challenges which will enhance the physics program. More broadly, the framework developed for Accelerator Real-time Edge AI Distributed Systems (READS) can be applied to future projects as the accelerator complex is upgraded for the PIP-II and DUNE era.
Highly distributed training of Deep Neural Networks (DNNs) on future compute platforms (offering 100 of TeraOps/s of computational capacity) is expected to be severely communication constrained. To overcome this limitation, new gradient compression techniques are needed that are computationally friendly, applicable to a wide variety of layers seen in Deep Neural Networks and adaptable to variations in network architectures as well as their hyper-parameters. In this paper we introduce a novel technique - the Adaptive Residual Gradient Compression (AdaComp) scheme. AdaComp is based on localized selection of gradient residues and automatically tunes the compression rate depending on local activity. We show excellent results on a wide spectrum of state of the art Deep Learning models in multiple domains (vision, speech, language), datasets (MNIST, CIFAR10, ImageNet, BN50, Shakespeare), optimizers (SGD with momentum, Adam) and network parameters (number of learners, minibatch-size etc.). Exploiting both sparsity and quantization, we demonstrate end-to-end compression rates of ~200X for fully-connected and recurrent layers, and ~40X for convolutional layers, without any noticeable degradation in model accuracies.
In many industry scale applications, large and resource consuming machine learning models reside in powerful cloud servers. At the same time, large amounts of input data are collected at the edge of cloud. The inference results are also communicated to users or passed to downstream tasks at the edge. The edge often consists of a large number of low-power devices. It is a big challenge to design industry products to support sophisticated deep model deployment and conduct model inference in an efficient manner so that the model accuracy remains high and the end-to-end latency is kept low. This paper describes the techniques and engineering practice behind Auto-Split, an edge-cloud collaborative prototype of Huawei Cloud. This patented technology is already validated on selected applications, is on its way for broader systematic edge-cloud application integration, and is being made available for public use as an automated pipeline service for end-to-end cloud-edge collaborative intelligence deployment. To the best of our knowledge, there is no existing industry product that provides the capability of Deep Neural Network (DNN) splitting.
We present Distributed Equivalent Substitution (DES) training, a novel distributed training framework for large-scale recommender systems with dynamic sparse features. DES introduces fully synchronous training to large-scale recommendation system for the first time by reducing communication, thus making the training of commercial recommender systems converge faster and reach better CTR. DES requires much less communication by substituting the weights-rich operators with the computationally equivalent sub-operators and aggregating partial results instead of transmitting the huge sparse weights directly through the network. Due to the use of synchronous training on large-scale Deep Learning Recommendation Models (DLRMs), DES achieves higher AUC(Area Under ROC). We successfully apply DES training on multiple popular DLRMs of industrial scenarios. Experiments show that our implementation outperforms the state-of-the-art PS-based training framework, achieving up to 68.7% communication savings and higher throughput compared to other PS-based recommender systems.
suggested questions
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
