Zobrazeno 1 - 10
of 4 678
pro vyhledávání: '"ANDREWS, DAVID"'
Transformer neural networks (TNN) excel in natural language processing (NLP), machine translation, and computer vision (CV) without relying on recurrent or convolutional layers. However, they have high computational and memory demands, particularly o
Externí odkaz:
http://arxiv.org/abs/2411.18148
Extent of the Magnetotail of Venus From the Solar Orbiter, Parker Solar Probe and BepiColombo Flybys
Autor:
Edberg, Niklas J. T., Andrews, David J., Boldú, J. Jordi, Dimmock, Andrew P., Khotyaintsev, Yuri V., Kim, Konstantin, Persson, Moa, Auster, Uli, Constantinescu, Dragos, Heyner, Daniel, Mieth, Johannes, Richter, Ingo, Curry, Shannon M., Hadid, Lina Z., Pisa, David, Sorriso-Valvo, Luca, Lester, Mark, Sánchez-Cano, Beatriz, Stergiopoulou, Katerina, Romanelli, Norberto, Fischer, David, Schmid, Daniel, Volwerk, Martin
Publikováno v:
Journal of Geophysical Research: Space Physics, 129, e2024JA032603
We analyze data from multiple flybys by the Solar Orbiter, BepiColombo, and Parker Solar Probe (PSP) missions to study the interaction between Venus' plasma environment and the solar wind forming the induced magnetosphere. Through examination of magn
Externí odkaz:
http://arxiv.org/abs/2410.21856
Autor:
Kabir, MD Arafat, Kamucheka, Tendayi, Fredricks, Nathaniel, Mandebi, Joel, Bakos, Jason, Huang, Miaoqing, Andrews, David
Many recent FPGA-based Processor-in-Memory (PIM) architectures have appeared with promises of impressive levels of parallelism but with performance that falls short of expectations due to reduced maximum clock frequencies, an inability to scale proce
Externí odkaz:
http://arxiv.org/abs/2410.07546
Autor:
Kabir, MD Arafat, Kamucheka, Tendayi, Fredricks, Nathaniel, Mandebi, Joel, Bakos, Jason, Huang, Miaoqing, Andrews, David
Processor-in-Memory (PIM) overlays and new redesigned reconfigurable tile fabrics have been proposed to eliminate the von Neumann bottleneck and enable processing performance to scale with BRAM capacity. The performance of these FPGA-based PIM archit
Externí odkaz:
http://arxiv.org/abs/2410.04367
Autor:
Kabir, Ehsan, Kabir, Md. Arafat, Downey, Austin R. J., Bakos, Jason D., Andrews, David, Huang, Miaoqing
Transformer neural networks (TNNs) are being applied across a widening range of application domains, including natural language processing (NLP), machine translation, and computer vision (CV). Their popularity is largely attributed to the exceptional
Externí odkaz:
http://arxiv.org/abs/2409.14023
Transformer neural networks (TNN) have been widely utilized on a diverse range of applications, including natural language processing (NLP), machine translation, and computer vision (CV). Their widespread adoption has been primarily driven by the exc
Externí odkaz:
http://arxiv.org/abs/2409.13975