{"id":2644,"date":"2018-02-07T09:14:11","date_gmt":"2018-02-07T12:14:11","guid":{"rendered":"https:\/\/www.nachodelatorre.com.ar\/mosconi\/?p=2644"},"modified":"2018-02-07T09:14:11","modified_gmt":"2018-02-07T12:14:11","slug":"procesadores-de-aprendizaje-profundo-para-dispositivos-inteligentes-de-iot","status":"publish","type":"post","link":"https:\/\/www.fie.undef.edu.ar\/ceptm\/?p=2644","title":{"rendered":"Procesadores de aprendizaje profundo para dispositivos inteligentes de IoT"},"content":{"rendered":"

En solo unos pocos a\u00f1os, Inteligencia Artificial (AI) \/ Aprendizaje Profundo (Deep Learning (DL) \/ Reinforcement Learning (RL) \/ Machine Learning (ML) se han convertido en herramientas importantes para muchas industrias y ahora contin\u00faan en un r\u00e1pido ciclo de innovaci\u00f3n.<\/p>\n

In the past few years, the Artificial Intelligence field has entered a high growth phase, driven largely by advancements in Machine Learning methodologies like Deep Learning (DL) and Reinforcement Learning (RL). Combinations of those techniques demonstrate unprecedented performance in solving a wide range of problems, from\u00a0playing Go at super-human<\/a>\u00a0level to diagnosing cancer like a specialist.<\/p>\n

In our previous blogs,\u00a0Intelligent IoT and Fog Computing Trends<\/a>\u00a0and\u00a0The Rise of Ubiquitous Computer Vision In IoT<\/a>, we talked about some interesting use cases of DL in IoT. The applications will be both broad and deep. They are going to fuel the demand for new breeds of processors in coming decades.<\/p>\n

Deep Learning Workflow Overview<\/h2>\n

DL\/RL innovations are happening at an astonishing pace (thousands of papers with new algorithms are presented in numerous\u00a0AI related conferences<\/a>\u00a0every year). Though it is premature to predict the final winning solutions, hardware companies are racing to build processors, tools, and frameworks. They are trying to identify pain points and bottlenecks in DL workflows (Fig. 1), leveraging years of experience of researchers.<\/p>\n

Deep Learning Workflow<\/h2>\n
\"\"
Fig. 1: Basic Deep Learning Workflow<\/figcaption><\/figure>\n

Platforms For Training DL Models<\/h2>\n

Let\u2019s start with training platforms. Graphical Processing Units (GPU) based systems are usually the choice for training advanced DL models. Nvidia has long realized the advantages of using GPU for general purpose high performance computing.<\/p>\n

GPU has hundreds of compute cores that support a large number of hardware threads and high throughput floating point computations. Nvidia developed Compute Unified Device Architecture (CUDA) programming framework to make GPU friendly for scientists and machine learning experts to use.<\/p>\n

CUDA toolchain has improved overtime, providing researchers a flexible and friendly way to realize highly complex algorithms. A few years ago, Nvidia aptly identified the DL opportunity and persistently developed CUDA support for most of DL operations. Standard frameworks like Caffe, Torch, and Tensorflow all support CUDA.<\/p>\n

In cloud services like AWS, developers have a choice between using CPU or GPU (more specifically Nvidia GPU). Platform choice depends on the complexity of the neural networks, budget, and time. GPU based systems can usually cut the training time by several times over CPU but are more expensive (Fig. 2)<\/p>\n

\"\"
Fig. 2: AWS EC2 GPU Instances<\/figcaption><\/figure>\n

Alternatives to GPU \/ CPU<\/h3>\n

Alternatives are coming. Khronos proposed\u00a0OpenCL<\/a>\u00a0in 2009 which is an open standard for parallel computing on a wide range of hardwares like CPU, GPU, DSP or FPGA. It will enable other processors like AMD GPUs to enter DL training market, providing developers with more choices.<\/p>\n

However, it is still behind CUDA in DL library support. Hopefully, that situation will improve in the next few years. Intel is also developing processors customized for DL training through its Nervana acquisition.<\/p>\n

Competitive Landscape of DL Inference<\/h2>\n

DL inference is a very competitive market. Applications can be deployed at multiple levels, usually depending on the requirements of the use cases:<\/p>\n