Learn More
State-of-the-art deep neural networks (DNNs) have hundreds of millions of connections and are both computationally and memory intensive, making them difficult to deploy on embedded systems with limited hardware resources and power budgets. While custom hardware helps the computation, fetching weights from DRAM is two orders of magnitude more expensive than(More)
BACKGROUND Wearable devices are gaining increasing market attention; however, the monitoring accuracy and consistency of the devices remains unknown. OBJECTIVE The purpose of this study was to assess the consistency of the monitoring measurements of the latest wearable devices in the state of normal activities to provide advice to the industry and support(More)
Sorting is one of the most fundamental algorithms of computer science and is the bottleneck of many computing problems. Pursuing fast and stable sorting of large scale has been the optimization goal of many applications. CPU-only based software methods often failed to tap the potential of computing-intensive sorting algorithms with parallel traits. It(More)
The ray tracing algorithm is well-known for its ability to generate photo-realistic rendering effects. Recent years have witnessed a renewed momentum in pushing it to real-time for better user experience. Today the construction of acceleration structures, e.g., kd-tree, has become the bottleneck of ray tracing. A dedicated hardware architecture, FastTree,(More)
Sparsity helps reduce the computational complexity of deep neural networks by skipping zeros. Taking advantage of sparsity is listed as a high priority in the next generation DNN accelerators such as TPU[1]. The structure of sparsity, i.e., the granularity of pruning, affects the efficiency of hardware accelerator design as well as the prediction accuracy.(More)
The rapid growth of mobile Internet services has triggered the explosion of Internet data traffic going through mobile networks. This brings a big challenge to mobile operators because their existing networks and deployment plan on mobile networks can hardly meet the demand. To solve this problem, alternative solutions are needed while Wi-Fi offloading is(More)
A fuzzy integral sliding mode control method based on the finite time stable theory and integral sliding mode control theory is proposed to solve a class of complex network problems. The complex networks are expressed by T-S fuzzy models with bounded approximation errors by using the approximation capability of T-S fuzzy models. The paper proposes fuzzy(More)