We've been using TensorFlow in daily research and engineering since it was released almost six months ago. We've learned a lot of things along the way. Because there are many subjective articles on TensorFlow and not enough helpful documentation, I've sprinkled in examples, tutorials, docs, and code snippets wherever possible. Community engagement is the most important thing. When it comes to machine learning, it is easy to focus on the tech (features, capabilities, benchmarks, etc).

There have been some remarkable developments lately in the world of artificial intelligence, from much publicized progress with self-driving cars to machines now composing Chopin imitations or just being really good at video games. Central to these advances are a number of tools around to help derive deep learning and other machine learning models, with Torch, Caffe, and Theano amongst those at the fore. However, since Google Brain went open source in November 2015 with their own framework, TensorFlow, we have seen the popularity of this software library skyrocket to be the most popular deep learning framework. Reasons include the wealth of support and documentation available, its production readiness, the ease of distributing calculations across a range of devices, and an excellent visualization tool: TensorBoard. Ultimately, TensorFlow manages to combine a comprehensive and flexible set of technical features with great ease of use.

Deep learning methods are achieving state-of-the-art results on challenging machine learning problems such as describing photos and translating text from one language to another. In this new laser-focused Ebook written in the friendly Machine Learning Mastery style that you're used to, finally cut through the math, research papers and patchwork descriptions about natural language processing. Using clear explanations, standard Python libraries and step-by-step tutorial lessons you will discover what natural language processing is, the promise of deep learning in the field, how to clean and prepare text data for modeling, and how to develop deep learning models for your own natural language processing projects. Click to jump straight to the packages. We are awash with text, from books, papers, blogs, tweets, news, and increasingly text from spoken utterances. Every day, I get questions asking how to develop machine learning models for text data. Working with text is hard as it requires drawing upon knowledge from diverse domains such as linguistics, machine learning, statistical natural language processing, and these days, deep learning. The problem with modeling text is that it is messy, and machine learning algorithms prefer well defined fixed-length inputs and outputs.

Classification of documents is a good NLP use case, but this work is moving further away from the goal of acquiring the semantics of documents rather than just "typing" it. This is where ML reach its limit. What if there are documents containing semantics that is totally new. These are common encounters in our everyday life as we need to process new technology, concept, and science. Practical use of ML related NLP work stops as we try to get down to the semantics of each document.

Joel Grus is a research engineer at the Allen Institute for Artificial Intelligence and the author of the bestselling O'Reilly book Data Science from Scratch: First Principles with Python. Previously he was a software engineer at Google and a data scientist at a variety of startups. He lives in Seattle, where he organizes various Data Science Happy Hours. This talk walks us through the works of Allen Institute for Artificial Intelligence and tells us about some of its projects. The main focus was on an AI based question answering system named Aristo and it's open source version Aristo-mini which utilizes various NLP and deep learning approach to answering science questions.

Deep Learning (DL) is part of the field of Machine Learning (ML). DL works by approximating a solution to a problem using neural networks. One of the nice properties of about neural networks is that they find patterns in the data (features) by themselves. This is opposed to having to tell your algorithm what to look for, as in the olde times. However, often this means the model starts with a blank state (unless we are transfer learning).

There's a buzz in AI circles around "capsule networks," a new variant on neural networks that backers say could simplify, cut the costs of, commoditize and, in the end, democratize how deep learning systems are taught to do what we want them to do. How can it do all this? Capsule networks hold out the hope of tacking one of the biggest problems in AI: radically reducing the amount of data (and compute) needed to train deep learning systems. This in turn means AI could become available to the broader market, no longer consigned to a few companies with mammoth compute resources and infinite volumes of data – i.e., the FANG* companies. In fact, a FANG company, Google, is the father of capsule networks.

Engineers at Purdue University in Lafayette, Indiana are developing a new system for keeping nuclear reactors safe with artificial intelligence (AI). In the paper published in the IEEE Transactions on Industrial Electronics journal, the researchers introduced a deep learning framework called a naïve Bayes-convolutional neural network that can effectively identify cracks in reactors by analyzing individual video frames. The method could potentially make safety inspections safer. "Regular inspection of nuclear power plant components is important to guarantee safe operations," Mohammad Jahanshahi, an assistant professor at Purdue's Lyles School of Civil Engineering, said in a press release. "However, current practice is time-consuming, tedious, and subjective and involves human technicians reviewing inspection videos to identify cracks in reactors."

I love solving puzzles and building things. Practicing data science gives me the opportunity to do both in equal measure. Like most data scientists, I came to the field indirectly. I started by studying robotics and human rehabilitation at MIT (MS '99, PhD '02), moved on to machine vision and machine learning at Sandia National Laboratories, then to predictive modeling of agriculture DuPont Pioneer, to cloud data science at Microsoft, and finally to satellite image processing at Facebook. In my spare time I like to rock climb, write robot learning algorithms, and go on walks with my wife and our dog, Reign of Terror.

Loquacious intelligent assistants have become a standard fixture of consumer devices, such as cell phones and smartwatches. These are harbingers of the accelerating osmosis of AI into everyday life. While charming, current implementations are pale imitations of what's coming. With most of the intelligence happening on cloud server farms, today's products are more like a ventriloquist's dummy, parroting responses from the real brains behind the curtain; smart, but limited. The emergence of Face ID, Apple's wondrous new biometric authentication system that uses facial recognition backed by an array of sensors and a new AI-accelerated iPhone SoC, marks the beginning of the second stage of embedded AI in which more of the intelligence happens on the device, independent of the cloud.