Abstract
The convolutional neural network (CNN) is one of the most used deep learning models for image detection and classification, due to its high accuracy when compared to other machine learning algorithms. CNNs achieve better results at the cost of higher computing and memory requirements. Inference of convolutional neural networks is therefore usually done in centralized high-performance platforms. However, many applications based on CNNs are migrating to edge devices near the source of data due to the unreliability of a transmission channel in exchanging data with a central server, the uncertainty about channel latency not tolerated by many applications, security and data privacy, etc. While advantageous, deep learning on edge is quite challenging because edge devices are usually limited in terms of performance, cost, and energy. Reconfigurable computing is being considered for inference on edge due to its high performance and energy efficiency while keeping a high hardware flexibility that allows for the easy adaption of the target computing platform to the CNN model. In this paper, we described the features of the most common CNNs, the capabilities of reconfigurable computing for running CNNs, the state-of-the-art of reconfigurable computing implementations proposed to run CNN models, as well as the trends and challenges for future edge reconfigurable platforms.
Highlights
Many machine learning applications are migrating from the cloud to the edge, close to where data is collected so as to overcome several negative issues associated with data computing on the cloud
Several different technologies are available for the execution of deep neural networks, namely reconfigurable devices, ASICs, GPUs, and CPUs, with different tradeoffs related to area, performance, cost, flexibility, energy, and others
Instead of a dense network model that applies generically to any input data type, a convolutional neural network (CNN) considers the spatial information between pixels, that is, the output of a neuron of the input layer is the result of the convolution between a small subset of the image and a kernel of weights
Summary
Many machine learning applications are migrating from the cloud to the edge, close to where data is collected so as to overcome several negative issues associated with data computing on the cloud (see Table 1). Deep neural networks (DNNs) are computationally demanding and memory hungry so it is a challenge to run these models in edge devices. Some approaches have minimized the size of neural networks and still keep the accuracy, like MobileNet [1], while others reduce the size or the number of parameters [2]. Another design option towards the implementation of deep neural networks in edge devices is the use of devices with better performance efficiency. Since the publication of AlexNet in 2012, many architectures have been proposed to accelerate deep neural networks [3,4] These works can be classified according within several dimensions (see Figure 1)
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
