… In the Figure 3(c), the blue vehicle behind DIA A3 ran away and the ego vehicle accelerated,
so our method predicted ego vehicle would insert into DIA A2, which means passing before …

… Learning (ML) models. An architecture has been designed and applied to a university's
satellite building, providing a flexible and adaptable network infrastructure for a smart building. …

… networks (BNNs) architecture play a vital role in the development of deep learning accelerator
for … We propose and present a reconfigurable BNN accelerator to improve the computing …

… is a promising hardware platform for accelerating deep neural networks (DNNs) thanks to
its … as a compelling ML acceleration platform due to its adaptability to frequent algorithm …

… Winograd algorithm and expands adaptability of Winograd for various convolution shapes,
which reduces computational complexity at algorithm level. The accelerator employs a high …

… device that accelerates convolutional neural networks. The … PEs communicate through a
scalable and adaptable network… objective of accelerating convolutional neural networks for …

… focus on accelerating SpMM using GPUs [7, 10–13]. However, accelerating SpMM suffers
from … Inspired by the prefix sum network, we propose the conditional reduction (Technique 4), …

… III-B3, it only takes 62.3 s to train such a network on our machine. The other nine DNNs
also … approach, revealing the excellent adaptability achieved by our designed optimization …

… Conventional neural network (CNN) accelerators use the time division multiplexing method,
and different network layers use the same accelerator, and their adaptability and resource …

… , the acceleration boost is large when training accuracy is high (our 528-250-125-10 network
trained on MNIST), but smaller when training accuracy is low (our 784-180-125-10 network …