On-device neural network training has emerged as a key enabler for privacy-preserving and personalized data processing at the edge. However, the deployment of Few-Shot Learning …
Sensors capture large volumes of data containing both critical and redundant information for neural network (NN) inference. However, limited bandwidth and energy efficiency at the …
To accelerate time consuming process for DNA alignment we propose a pre-mapping filtering method, called the Same Token Count (STC), that leverages the high parallelism of …
In this work, we propose a Vision Transformer (ViT) hardware accelerator that can achieve high utilization and high efficiency. Unlike the convolutional neural network (CNN) …
The key to the video super-resolution algorithm is to capture the information of adjacent frames to supplement the reconstruction of the current frame. It is necessary to apply the …
In this letter, we predict the locations as probability distributions for the tasks of image object detection. We adopt the Kullback-Leibler divergences as the regression losses to …
In this paper, we propose an edge-guided video super-resolution (EGVSR) network that utilizes the edge information of the image to effectively recover high-frequency details for …
This paper presents a model compression frame-work for both pruning and quantizing according to the channel distribution information. We apply the variational inference technique …
Convolutional neural network (CNN), one of the branches of deep neural networks, has been widely used in image recognition, natural language processing, and other related fields …
Recently, studies on single image super-resolution using Deep Convolutional Neural Networks (DCNN) have been demonstrated to have made outstanding progress over conventional …