Nick Bostrom, one of the leading voices on artificial intelligence, has singled out London research lab DeepMind as the company closest to developing a system that can mimic human-level artificial intelligence -- a target widely shared by those at the forefront of the AI industry. When asked who was leading the global AI race, Bostrom immediately responded with DeepMind. "Right now, I think here in London we have the DeepMind group who are, I think, the biggest [group] specifically focused on solving general intelligence," Bostrom told Business Insider at a breakfast meeting aboard the Sunbourn Yacht Hotel in East London on Wednesday. DeepMind, which employs approximately 250 people in King's Cross, was acquired by Google in 2014 for a reported £400 million. The organisation is perhaps best known for developing an AI agent that defeated the world champion of the ancient Chinese board, Go.

Nick Bostrom, one of the leading voices on artificial intelligence, has singled out London research lab DeepMind as the company closest to developing a system that can mimic human-level artificial intelligence -- a target widely shared by those at the forefront of the AI industry. When asked who was leading the global AI race, Bostrom immediately responded with DeepMind. "Right now, I think here in London we have the DeepMind group who are, I think, the biggest [group] specifically focused on solving general intelligence," Bostrom told Business Insider at a breakfast meeting aboard the Sunbourn Yacht Hotel in East London on Wednesday. DeepMind, which employs approximately 250 people in King's Cross, was acquired by Google in 2014 for a reported £400 million. The organisation is perhaps best known for developing an AI agent that defeated the world champion of the ancient Chinese board, Go.

Representing a dialog policy as a recurrent neural network (RNN) is attractive because it handles partial observability, infers a latent representation of state, and can be optimized with supervised learning (SL) or reinforcement learning (RL). For RL, a policy gradient approach is natural, but is sample inefficient. In this paper, we present 3 methods for reducing the number of dialogs required to optimize an RNN-based dialog policy with RL. The key idea is to maintain a second RNN which predicts the value of the current policy, and to apply experience replay to both networks. On two tasks, these methods reduce the number of dialogs/episodes required by about a third, vs. standard policy gradient methods.

During the dot-com boom, online clothing sales were predicted to grow to 40% -50% of total sales. Although online sales of some other kinds of merchandise, such as books, have reached 50% of the market in the past 15 years, the percentage of online clothing sales hovers around 20%. The difficulty in finding the correct size and fit is one of the primary reasons that consumers are reluctant to buy clothes online. And their concern is not groundless; sizing varies among clothing manufacturers, and it is difficult to ascertain fit from online images. Consequently, 30%-40% of online clothing purchases are returned.

Machine learning is a computer's ability to learn and adapt without being explicitly programmed. This is a widely useful technology that aids in banking, DNA sequencing, search engines, and myriad other applications. If this sounds like a career you'd be interested in, then you'll want to learn all there is to know about machine learning, and you'll want to start from the groun up. Luckily, Windows Central Digital Offers has three awesome course bundles that'll get you up and running and on your way to programming machine learning and AI -- all for $120! This bundle takes you from the basics of machine learning to some advanced techniques, as well as learning to code with Python.

We've paid a lot of attention lately to Convolutional Neural Nets (CNNs) as the cornerstone of 2nd gen NNs and spent some time on Spiking Neural Nets (SNNs) as the most likely path forward to 3rd gen, but we'd really be remiss if we didn't stop to recognize Recurrent Neural Nets (RNNs). Because RNNs are solid performers in the 2nd gen NN world and perform many tasks much better than CNNs. These include speech-to-text, language translation, and even automated captioning for images. By count, there are probably more applications for RNNs than for CNNs. On one scale RNNs have much more in common with the larger family of NNs than do CNNs which have very unique architecture.

Thank you for joining us on December 8th at Travana in San Francisco - we had a great time, hope you did too! Sergei described the rise of bots and what it means for the travel space, and where the opportunities are. He talked about how Deep Learning is utilized for building smart conversational bots and compared them with less intelligent bots that rely on on-screen buttons. He opened up the questioned of whether AI would change the way users learn directly from reviews. TTC also announced the exciting Travel Tech Con 2017 that will be held at Treasure Island in May 2017!

Starred articles are candidates for the picture of the week. A comprehensive list of all past resources is found here. We are in the process of automatically categorizing them using indexation and automated tagging algorithms. How Machine Learning Can Help Increase Cybersecurity Extraordinary Link Between Deep Neural Networks and the Nature of t... Every Data Science Interview Boiled Down To Five Basic Questions How fog computing pushes IoT intelligence to the edge Machine learning's next trick is generating videos from photos Machine learning just got more human with Google's RankBrain Why do traffic jams sometimes form for no reason? Data Mining Reveals the Six Basic Emotional Arcs of Storytelling visualization of people commuting to work to Manhattan * Machine Learning over 1M hotel reviews finds interesting insights Illegal in Massachusetts: Asking Your Salary in a Job Interview Data Science Tops List of Fields with Massive Potential Statistics Denial, Applied Statistics Is A Way Of Thinking, Not Jus... EU citizens might get a'right to explanation' about the decisions ... 3 blossoming fields of study with massive potential - #1 is data science Machine learning helps scientists discover new materials Fighting ISIS With an Algorithm, Physicists Try to Predict Attacks How to navigate a city without using any street names Google given access to healthcare data of up to 1.6 million patients Machine learning has boosted Google's translation capabilities to n... Machine learning's next trick is generating videos from photos Machine learning just got more human with Google's RankBrain Why do traffic jams sometimes form for no reason?

In a previous article, we had an overview of the applications of Deep Learning and touched upon some basic points to consider while creating a Deep Learning model. We also had an overview of what it is and methods to get started with deep learning. In this article, we jump straight into creating an anomaly detection model using Deep Learning and anomaly package from H2O. Readers who don't know what it is can view it as anything that occurs unexpected and is a rare event. It is a deviation from the standard pattern and does not confirm to the usual behavior of the data. Let's say we work in a steel manufacturing industry, and we see the quality of the steel suddenly drops down below the permissible limits. This is an anomaly; if not detected and resolved soon will cost the organization millions.

When Ray Kurzweil met with Google CEO Larry Page last July, he wasn't looking for a job. A respected inventor who's become a machine-intelligence futurist, Kurzweil wanted to discuss his upcoming book How to Create a Mind. He told Page, who had read an early draft, that he wanted to start a company to develop his ideas about how to build a truly intelligent computer: one that could understand language and then make inferences and decisions on its own. It quickly became obvious that such an effort would require nothing less than Google-scale data and computing power. "I could try to give you some access to it," Page told Kurzweil.