In the past 10 years, the best-performing artificial-intelligence systems -- such as the speech recognizers on smartphones or Google's latest automatic translator -- have resulted from a technique called "deep learning." Deep learning is in fact a new name for an approach to artificial intelligence called neural networks, which have been going in and out of fashion for more than 70 years. Neural networks were first proposed in 1944 by Warren McCullough and Walter Pitts, two University of Chicago researchers who moved to MIT in 1952 as founding members of what's sometimes called the first cognitive science department. Neural nets were a major area of research in both neuroscience and computer science until 1969, when, according to computer science lore, they were killed off by the MIT mathematicians Marvin Minsky and Seymour Papert, who a year later would become co-directors of the new MIT Artificial Intelligence Laboratory. The technique then enjoyed a resurgence in the 1980s, fell into eclipse again in the first decade of the new century, and has returned like gangbusters in the second, fueled largely by the increased processing power of graphics chips. "There's this idea that ideas in science are a bit like epidemics of viruses," says Tomaso Poggio, the Eugene McDermott Professor of Brain and Cognitive Sciences at MIT, an investigator at MIT's McGovern Institute for Brain Research, and director of MIT's Center for Brains, Minds, and Machines.

This article is a gentle introduction to attentional and memory-based interfaces in deep neural architectures, using TensorFlow. Incorporating attention mechanisms is very simple and can offer transparency and interpretability to our complex models. We conclude with extensions and caveats of the interfaces. As you read the article, please access all of the code on GitHub and view the IPython notebook here; all code is compatible with TensorFlow version 1.0. The intended audience for this notebook are developers and researchers who have some basic understanding of TensorFlow and fundamental deep learning concepts.

Click on the blue word above the attention of the medical profession, every day there are material! When it comes to Dr. Siddartha (Siddhartha Mukherjee), perhaps a lot of people are unfamiliar with his name. But a lot of people are familiar with his two book, "the king of all diseases: cancer," and "genes: Intimate History.". The former allows Dr. Mukherjee to get a non fiction Pulitzer prize, while the latter was recommended as the best book of 2016 by Mr. Bill Gate. Recently, Dr. Mukherjee in "New York guest" (The New Yorker) published a long article, the unique perspective of a doctor, artificial intelligence survey in recent years the impact for medicine. The seven story is he in this long article record, outlines the future doctors and artificial intelligence, harmonious coexistence. The author Dr. Mukherjee is a doctor, but also a good writer. One night in November 2016, a 54 year old woman in New York, Bronx (Bronx) was sent to the emergency room at the Columbia University (Columbia University) medical center because of a severe headache. She told the emergency room doctor that his vision was blurred and his left hand was numb. The doctor arranged for CT. A few months later, on January, one of the 4 radiologists huddled in front of a computer on the third floor of the hospital, the room was dark and windowless, with only the screen light, which seemed to be filtered by the sea. She's training them to read CT. "Once the brain shows death and gray, it's easy to diagnose a stroke," Dr. Lignelli-Dipple said. The key is to diagnose a stroke before most nerve cells die." A stroke is usually caused by a blockage or bleeding of the blood vessel. The radiologist has about 45 minutes of window time so that the doctor can intervene in time to dissolve the clot. "Imagine you're in the emergency room right now," continued Dr. Lignelli-Dipple. "Every minute, a part of the brain dies.

About halfway through a particularly tense game of Go held in Seoul, South Korea, between Lee Sedol, one of the best players of all time, and AlphaGo, an artificial intelligence created by Google, the AI program made a mysterious move that demonstrated an unnerving edge over its human opponent. On move 37, AlphaGo chose to put a black stone in what seemed, at first, like a ridiculous position. It looked certain to give up substantial territory--a rookie mistake in a game that is all about controlling the space on the board. Two television commentators wondered if they had misread the move or if the machine had malfunctioned somehow. In fact, contrary to any conventional wisdom, move 37 would enable AlphaGo to build a formidable foundation in the center of the board. The Google program had effectively won the game using a move that no human would've come up with. AlphaGo's victory is particularly impressive because the ancient game of Go is often looked at as a test of intuitive intelligence. The rules are quite simple. Two players take turns putting black or white stones at the intersection of horizontal and vertical lines on a board, trying to surround their opponent's pieces and remove them from play.

NVIDIA TensorRT is a high-performance deep learning inference library for production environments. Power efficiency and speed of response are two key metrics for deployed deep learning applications, because they directly affect the user experience and the cost of the service provided. Tensor RT automatically optimizes trained neural networks for run-time performance, delivering up to 16x higher energy efficiency (performance per watt) on a Tesla P100 GPU compared to common CPU-only deep learning inference systems (see Figure 1). Figure 2 shows the performance of NVIDIA Tesla P100 and K80 running inference using TensorRT with the relatively complex GoogLenet neural network architecture. In this post we will show you how you can use Tensor RT to get the best efficiency and performance out of your trained deep neural network on a GPU-based deployment platform.

From Siri to Google Translate, deep neural networks have enabled breakthroughs in machine understanding of natural language. Most of these models treat language as a flat sequence of words or characters, and use a kind of model called a recurrent neural network (RNN) to process this sequence. But many linguists think that language is best understood as a hierarchical tree of phrases, so a significant amount of research has gone into deep learning models known as recursive neural networks that take this structure into account. While these models are notoriously hard to implement and inefficient to run, a brand new deep learning framework called PyTorch makes these and other complex natural language processing models a lot easier. While recursive neural networks are a good demonstration of PyTorch's flexibility, it is also a fully-featured framework for all kinds of deep learning with particularly strong support for computer vision.

There is increasing interplay between the worlds of machine learning and high performance computing (HPC). This began with a shared hardware and software story since many supercomputing tricks of the trade play well into deep learning, but as we look to next generation machines, the bond keeps tightening. Many supercomputing sites are figuring out how to work deep learning into their existing workflows, either as a pre- or post-processing step, while some research areas might do away with traditional supercomputing simulations altogether eventually. While these massive machines were designed with simulations in mind, the strongest supers have architectures that parallel the unique requirements of training and inference workloads. One such system in the U.S. is the future Summit supercomputer coming to Oak Ridge National Lab later this year, but many of the other architectures that are especially sporting for machine learning are in China and Japan--and feature non-standard processing elements.

The latest research from DeepMind is proving how inspired the idea to model neural networks of the human mind truly was. The strength of the association between the human brains and and their computational models is revealing weaknesses in our own minds and teaching us how to overcome them. Google's engineers, inspired by neuroscience, have created an Artificial Intelligence (AI) using an Artificial Neural Network that can hang onto knowledge as it moves from task to task, spinning the straw of raw memory into gold that stays with the program, forming long-term experiences. And while human minds do this after a fashion, we are not as adept at discerning what is important. You may recall the song lyrics you heard when you first rode a bicycle as well as you recall important information from your career successes.

As of February 2017, there were 106 artificial intelligence startups in healthcare, according to a CB Insights report, and 70 of them launched last year. Here are 13 healthcare AI startups that have raised $25 million or more, listed along with their investors. Flatiron's platform connects community practices and cancer centers on a common technology infrastructure to address healthcare challenges with the goal of powering a national benchmarking and research network for cancer care. The company provides its platform to more than 265 community cancer clinics and three major academic research centers. Welltok's CafeWell Health Optimization Platform is designed to connect consumers with benefits, resources and rewards for personalized healthcare plans.

Jupyter is an excellent interface for executing deep learning development and training. In fact, many of the tutorials that help you get started with deep learning frameworks use Jupyter notebooks because of the ability to provide small blocks of commented, executable code with easy reproducibility, ability to display intermediate feature extraction information, and useful metrics and console output. However, one of the major challenges with using Jupyter and with deep learning more generally is the complexity of context switching between prototyping with CPUs and accelerated training with GPUs. At Bitfusion, we've developed custom kernels coupled with network-attached Elastic GPUs to make it quick and easy to switch from CPUs to GPUs and back again with only a couple clicks.