Artificial Intelligence (AI) is making progress in great strides, or at least it appears so! Almost no week passes by without some major announcements of new challenges solved by AI technology or new products powered by AI. Indeed many quantifiable factors attest an unprecedented level of activity: capital investments, number of academic papers, number of products involving AI technology, they all are on a steep rise in the past five years. Computers are already very capable at some specialized tasks that require reasoning and other abilities that we typically associate with intelligence. For example, computers can play a decent game of chess or can help us order our holiday photos. Despite this genuine progress, we are still a long way from human level intelligence because our best artificial intelligence systems are not general purpose. They cannot quickly adapt to novel tasks the way most humans can do.

In the previous blog post Using Keras and Deep Q-Network to Play FlappyBird we demonstrate using Deep Q-Network to play FlappyBird. However, a big limitation of Deep Q-Network is that the outputs/actions are discrete while the action like steering are continuous in car racing. An obvious approach to adapt DQN to continuous domains is to simply discretize the action space. However, we encounter the "curse of dimensionality" problem. For example, if you discretize the steering wheel from -90 to 90 degrees in 5 degrees each and acceleration from 0km to 300km in 5km each, your output combinations will be 36 steering states times 60 velocity states which equals to 2160 possible combinations. The situation will be worse when you want to build robots to perform something very specialized, such as brain surgery that requires fine control of actions and naive discretization will not able to achieve the required precision to do the operations.

Learning incorporates a broad range of complex procedures. Machine learning (ML) is a subdivision of artificial intelligence based on the biological learning process. The ML approach deals with the design of algorithms to learn from machine readable data. ML covers main domains such as data mining, difficult-to-program applications, and software applications. It is a collection of a variety of algorithms (e.g.

Nina Zumel recently mentioned the use of Laplace noise in "count codes" by Misha Bilenko (see here and here) as a known method to break the overfit bias that comes from using the same data to design impact codes and fit a next level model. It is a fascinating method inspired by differential privacy methods, that Nina and I respect but don't actually use in production. Please read on for my discussion of some of the limitations of the technique, and how we solve the problem for impact coding (also called "effects codes"), and a worked example in R.We define a nested model as any model where the results of a sub-model are used as inputs for a later model. Common examples include variable preparation, ensemble methods, super-learning, and stacking. Nested models are very common in machine learning.

The lab they have created has also had some big wins, not only for them, but also for the networks on which their phones run. For instance, their work has helped carriers across the world to reduce their pay-as-you-go customer churn rate from 10% to 6%. They have been working on a machine learning driven network control system, which will achieve automated network traffic control. 'Tests indicate that Network Mind is up to 500% more efficient in realizing KPIs such as task completion or policy generation compared to existing template or heuristic algorithm-based optimization methods. Network Mind is also over 50 times more efficient when analyzing paths of large optical networks, which has the potential to reduce the time it takes to analyze use cases such as optical network failure prevention from 5 hours to as little as 6 minutes.'

Artificial intelligence is the key to the future of online retail, business bosses have said, providing a crucial way to help shoppers find what they want. Alex Baldock, chief executive of Shop Direct, which runs very.co.uk and Littlewoods, told the Telegraph Festival of Business in London that artificial intelligence was the company's "big bet". "You have three seconds to seize the shopper's attention - it's called thumb stopping, the three-second audition," Mr Baldock said. "That's where personalisation comes in." Shop Direct is owned by Sir David and Sir Frederick Barclay, proprietors of Telegraph Media Group, the publisher of the Daily Telegraph.

What AI Can and Can't Do Many executives ask me what artificial intelligence can do. They want to know how it will disrupt their industry and how they can use it to reinvent their own companies. But lately the media has sometimes painted an unrealistic picture of the powers of AI. (Perhaps soon it will take over the world!) AI is already transforming web search, advertising, e-commerce, finance, logistics, media, and more. As the founding lead of the Google Brain team, former director of the Stanford Artificial Intelligence Laboratory, and now overall lead of Baidu's AI team of some 1,200 people, I've been privileged to nurture many of the world's leading AI groups and have built many AI products that are used by hundreds of millions of people. Having seen AI's impact, I can say: AI will transform many industries.

Many predict artificial intelligence (AI) will someday cure cancer, clean up our environment, drastically improve our cities, and send us to Mars, among a host of other accomplishments. However, with the potential benefits that inspire oohs and aahs also comes risk. Before there was Siri, there was HAL. And in the epic science fiction movie "2001: A Space Odyssey," when the poor astronaut Dave Bowman asks the supercomputer HAL 9000 to let him back inside the spacecraft, HAL merely responds, "I'm sorry Dave, I'm afraid I can't do that." Granted, this was a movie--albeit a brilliant one--but there are still real life concerns about what exactly are artificial intelligence best practices and a need for guidelines.

I. Acceptance of this Agreement By downloading or otherwise accessing the Materials, Downloader represents his/her acceptance of the terms of this Agreement. Modification of this Agreement Users may modify the terms of this Agreement at any time. However, any modifications to this Agreement will only be effective for downloads subsequent to such modification. No modifications will supersede any previous terms that were in effect at the time of the Downloader's download. Use of the Materials Use of the Materials include but are not limited to viewing parts or the whole of the content included in the Materials; comparing data or content from the Materials with data or content in other Materials; verifying research results with the content included in the Materials; and extracting and/or appropriating any part of the content included in the Materials for use in other projects, publications, research, or other related work products.

For a while (at least several months since many people began to implement it with Python and/or Theano, PyLearn2 or something like that), nearly I've given up practicing Deep Learning with R and I've felt I was left alone much further away from advanced technology… But now we have a great masterpiece: {h2o}, an implementation of H2O framework in R. I believe {h2o} is the easiest way of applying Deep Learning technique to our own datasets because we don't have to even write any code scripts but only to specify some of its parameters. That is, using {h2o} we are free from complicated codes; we can only focus on its underlying essences and theories. With using {h2o} on R, in principle we can implement "Deep Belief Net", that is the original version of Deep Learning*1. I know it's already not the state-of-the-art style of Deep Learning, but it must be helpful for understanding how Deep Learning works on actual datasets. Please remember a previous post of this blog that argues about how decision boundaries tell us how each classifier works in terms of overfitting or generalization, if you already read this blog.