Who has the advantage in artificial intelligence -- big companies, startups, or academia? Perhaps all three, especially as they work together when it comes to fields like this. One thing is clear though: A.I. and deep learning is where it's at. And that's why this year's newly anointed Andreessen Horowitz Distinguished Visiting Professor of Computer Science is Fei-Fei Li [who publishes under Li Fei-Fei], associate professor at Stanford University. Bridging entrepreneurs across academia and industry, we began the a16z Professor-in-Residence program just a couple years ago (most recently with Dan Boneh and beginning with Vijay Pande).

Much of the buzz around artificial intelligence has surrounded companies focused on general purpose artificial intelligence, as opposed to companies applying AI algorithms like machine learning and deep learning in specific industries. "From healthcare to finance to e-commerce, we're focused on changing people's lives." Taking a look at the core AI companies like Sentient -- companies focused on general-purpose AI applicable across a variety of industries -- a few have already received 50M or more in equity funding. These include Sentient, Ayasdi, Digital Reasoning, Vicarious (whose backers include angels Jeff Bezos, Elon Musk, and Marc Benioff) as well as DataRobot. We used CB Insights data to look at the most well-funded core AI companies, below.

The book is more like a quick compilation of a college student's note. Concepts are presented in a relatively isolated manner; connections between concepts are, for the large part, missing. Furthermore, if the materials presented are rather shallow like in this book, readers will expect to see a strong emphasis on, or hands on exercises of, practical applications. But this book doesn't seem to help much in that regard either, despite what the book claims. The book does give introduction to a bunch of models, which can be useful for a beginner. But at least this edition I wouldn't suggest any one to buy because of poor editing.

Neural Machine Translation (MT) has reached state-of-the-art results. However, one of the main challenges that neural MT still faces is dealing with very large vocabularies and morphologically rich languages. In this paper, we propose a neural MT system using character-based embeddings in combination with convolutional and highway layers to replace the standard lookup-based word representations. The resulting unlimited-vocabulary and affix-aware source word embeddings are tested in a state-of-the-art neural MT based on an attention-based bidirectional recurrent neural network. The proposed MT scheme provides improved results even when the source language is not morphologically rich. Improvements up to 3 BLEU points are obtained in the German-English WMT task.

Machine learning has undergone a flurry of progress recently thanks to the growth of big data, fast hardware, and clever algorithms. And with tools such as Spark and Hadoop, the JVM is no stranger to machine learning, using tools including h2o, dl4j and other Spark-based libraries. With these tools, developers can harness the power of distributed hardware and deep learning to discover new and untapped patterns and relationships in big data. In this session, learn how to train a classifier to recognize handwritten digits and how you can build your own models using open source data sets. No prior experience is required.

Towards making systems forget with machine unlearning.

OpenAI, a nonprofit created by Elon Musk and other tech entrepreneurs to make fundamental breakthroughs in artificial intelligence, has said that one of its big goals will be teaching robots to do the laundry and other household chores. OpenAI doesn't want to make robot hardware itself but, rather, to supply the brains for off-the-shelf bots. You might think that learning to fold underpants is a modest goal, but such dexterity and adaptability is one of the grand challenges of robotics. It also fits with the organization's stated objective to "advance digital intelligence in the way that is most likely to benefit humanity as a whole." Applying the sort of machine-learning techniques OpenAI is working on to robotics should, in fact, have huge practical benefits, and it will be a necessary component of any more general form of artificial intelligence.

Whether it's Deep Blue beating Garry Kasparov at chess, Watson defeating Ken Jennings at Jeopardy!, or Google DeepMind's AlphaGO besting Lee Sedo at Go, artificial intelligence can't be underestimated when it comes to taking on the champions and winning. That's because a new AI system called ALPHA -- developed by recent University of Cincinnati doctoral graduate Nick Ernest, now CEO of Psibernetix -- recently defeated retired United States Air Force Colonel Gene Lee in an air combat simulator. Not only did Colonel Lee, who has extensive aerial combat experience as an instructor, fail to kill ALPHA's aircraft during combat, he was also repeatedly shot out of the air by the bot. According to Lee, ALPHA is "the most aggressive, responsive, dynamic and credible AI I've seen to date." "ALPHA is an incredibly difficult opponent to face," Psibernetix CEO Nick Ernest tells Digital Trends. "Even flying against other pilots when ALPHA has severe handicaps to a number of its systems -- including speed, turning, missile capability and sensors -- it is able to win.

Rao goes on, "The production release of fully autonomous cars is probably at least five years away still, as these machines are not nearly safe enough for widespread consumer use. Google's self-driving cars still make mistakes, like getting confused by cyclists on fixed-gear bikes at stop signs. Tesla's Autopilot has run into trouble when driving on local streets instead of highways. In fact, there are an unlimited number of such corner cases that autonomous vehicles must respond to, and many still need to be discovered and factored in. Only when a sufficient number of scenarios have been addressed will autonomous cars be considered'safe enough.' As Tesla recently blogged: 'Getting [an autonomous car] to be 99% correct is relatively easy, but getting it to be 99.9999% Making mistakes at 70 mph would be highly problematic'."