The Next Generation of Deep Learning Hardware: Analog Computing
| العنوان: | The Next Generation of Deep Learning Hardware: Analog Computing |
|---|---|
| المؤلفون: | Tayfun Gokmen, Ruchir Puri, Wilfried Haensch |
| المصدر: | Proceedings of the IEEE. 107:108-122 |
| بيانات النشر: | Institute of Electrical and Electronics Engineers (IEEE), 2019. |
| سنة النشر: | 2019 |
| مصطلحات موضوعية: | Artificial neural network, business.industry, Computer science, Deep learning, Analog computer, Matrix multiplication, Rendering (computer graphics), law.invention, Non-volatile memory, Computer engineering, law, Artificial intelligence, Electrical and Electronic Engineering, Graphics, Mathematical structure, business |
| الوصف: | Initially developed for gaming and 3-D rendering, graphics processing units (GPUs) were recognized to be a good fit to accelerate deep learning training. Its simple mathematical structure can easily be parallelized and can therefore take advantage of GPUs in a natural way. Further progress in compute efficiency for deep learning training can be made by exploiting the more random and approximate nature of deep learning work flows. In the digital space that means to trade off numerical precision for accuracy at the benefit of compute efficiency. It also opens the possibility to revisit analog computing, which is intrinsically noisy, to execute the matrix operations for deep learning in constant time on arrays of nonvolatile memories. To take full advantage of this in-memory compute paradigm, current nonvolatile memory materials are of limited use. A detailed analysis and design guidelines how these materials need to be reengineered for optimal performance in the deep learning space shows a strong deviation from the materials used in memory applications. |
| تدمد: | 1558-2256 0018-9219 |
| DOI: | 10.1109/jproc.2018.2871057 |
| URL الوصول: | https://explore.openaire.eu/search/publication?articleId=doi_________::ea1b2bfb8ac5b4dbe989de102bbbf622 https://doi.org/10.1109/jproc.2018.2871057 |
| Rights: | CLOSED |
| رقم الانضمام: | edsair.doi...........ea1b2bfb8ac5b4dbe989de102bbbf622 |
| قاعدة البيانات: | OpenAIRE |
| تدمد: | 15582256 00189219 |
|---|---|
| DOI: | 10.1109/jproc.2018.2871057 |