The 1983 movie "War Games" has a memorable climax where the supercomputer known as WOPR (War Operation Plan Response) is asked to train on itself to discover the concept of an un-winnable game. The character played by Mathew Broderick asks "Is there any way that it can play itself?" The solution is the same, set the number of players to zero (i.e. There is plenty to digest about this latest breakthrough in Deep Learning technology. DeepMind authors use the term "self-play reinforcement learning". As I remarked in the piece about "Tribes of AI", DeepMind is particularly fond of their Reinforcement Learning (RL) approach.

Artificial intelligence: everybody is talking about it, and the as-of-yet unrealized possibilities of the technology are fueling a renaissance in the hardware and software industry. Hardware and software companies -- including Intel, NVidia, Google, IBM, Microsoft, Facebook, Qualcomm, ARM and many others -- are racing to build the next AI hardware platform or fighting to maintain their lead. AI, and deep learning (a sub-field of neural networks) in particular is an inherently non-Von Neumann process, and the prospect of having a processor more closely tailored to the specific needs of neural networks is appealing. But, I like to think before acting, especially before diving into a potentially very expensive hardware project. Should the AI industry build a specialized deep learning chip, and, if so, what should it look like?

Silicon Valley's start-ups have always had a recruiting advantage over the industry's giants: Take a chance on us and we'll give you an ownership stake that could make you rich if the company is successful. Now the tech industry's race to embrace artificial intelligence may render that advantage moot -- at least for the few prospective employees who know a lot about A.I. Tech's biggest companies are placing huge bets on artificial intelligence, banking on things ranging from face-scanning smartphones and conversational coffee-table gadgets to computerized health care and autonomous vehicles. As they chase this future, they are doling out salaries that are startling even in an industry that has never been shy about lavishing a fortune on its top talent. Typical A.I. specialists, including both Ph.D.s fresh out of school and people with less education and just a few years of experience, can be paid from $300,000 to $500,000 a year or more in salary and company stock, according to nine people who work for major tech companies or have entertained job offers from them. All of them requested anonymity because they did not want to damage their professional prospects.

Distributed, in-memory database enables correlation of customer profile, buying behaviour, sentiment and shopping trends based on cross-industry data sources in sub-seconds JAKARTA, Oct 24, 2017 - (ACN Newswire) - OVO, a member of Lippo Group Digital, has chosen Kinetica, provider of the world's fastest GPU-accelerated analytics database, and NVIDIA to spur innovation in big data and analytics that fully utilize a 360-degree customer view and derive real-time analytics and insights on shopping trends and the digital lifestyle. Lippo Group is a prominent conglomerate with significant investments in digital technologies, education, financial services, healthcare, hospitality, media, IT, telecommunications, real estate, entertainment and retail. Aggressively investing in big data and analytics technology, OVO is Lippo Group Digital's concierge platform, integrating mobile payment, loyalty points, and exclusive priority deals. With the support from Kinetica and NVIDIA, Lippo Group Digital will be the first enterprise to integrate an AI, in-memory, GPU database in Indonesia. Lippo Group Digital's investment in big data and analytics technology aims to consolidate all their customer data that are being generated by all transactional systems from various subsidiaries into a centralized analytics platform.

NVIDIA (NASDAQ:NVDA) is using its GPUs (graphics processing units) to make a big splash in the field of artificial intelligence (AI). The graphics specialist is leveraging the immense computational capacities of its GPUs, which can perform a huge number of mathematical calculations in a parallel manner, for enabling AI across several applications. NVIDIA has started witnessing terrific growth in some of its businesses that rely on AI. For instance, NVIDIA's automotive business increased almost 20% year over year in the last reported quarter; the company's DRIVE PX2 autonomous vehicle platform uses AI to help cars drive themselves. But this is just one of the many areas where NVIDIA is applying AI to boost its business.

This paper describes a novel text-to-speech (TTS) technique based on deep convolutional neural networks (CNN), without any recurrent units. Recurrent neural network (RNN) has been a standard technique to model sequential data recently, and this technique has been used in some cutting-edge neural TTS techniques. However, training RNN component often requires a very powerful computer, or very long time typically several days or weeks. Recent other studies, on the other hand, have shown that CNN-based sequence synthesis can be much faster than RNN-based techniques, because of high parallelizability. The objective of this paper is to show an alternative neural TTS system, based only on CNN, that can alleviate these economic costs of training. In our experiment, the proposed Deep Convolutional TTS can be sufficiently trained only in a night (15 hours), using an ordinary gaming PC equipped with two GPUs, while the quality of the synthesized speech was almost acceptable.

Transfer learning using deep neural networks as feature extractors has become increasingly popular over the past few years. It allows to obtain state-of-the-art accuracy on datasets too small to train a deep neural network on its own, and it provides cutting edge descriptors that, combined with nonparametric learning methods, allow rapid and flexible deployment of performing solutions in computationally restricted settings. In this paper, we are interested in showing that the features extracted using deep neural networks have specific properties which can be used to improve accuracy of downstream nonparametric learning methods. Namely, we demonstrate that for some distributions where information is embedded in a few coordinates, segmenting feature vectors can lead to better accuracy. We show how this model can be applied to real datasets by performing experiments using three mainstream deep neural network feature extractors and four databases, in vision and audio.

DeConvNet, Guided BackProp, LRP, were invented to better understand deep neural networks. We show that these methods do not produce the theoretically correct explanation for a linear model. Yet they are used on multi-layer networks with millions of parameters. This is a cause for concern since linear models are simple neural networks. We argue that explanation methods for neural nets should work reliably in the limit of simplicity, the linear models. Based on our analysis of linear models we propose a generalization that yields two explanation techniques (PatternNet and PatternAttribution) that are theoretically sound for linear models and produce improved explanations for deep networks.

Deep convolutional neural networks have led to breakthrough results in numerous practical machine learning tasks such as classification of images in the ImageNet data set, control-policy-learning to play Atari games or the board game Go, and image captioning. Many of these applications first perform feature extraction and then feed the results thereof into a trainable classifier. The mathematical analysis of deep convolutional neural networks for feature extraction was initiated by Mallat, 2012. Specifically, Mallat considered so-called scattering networks based on a wavelet transform followed by the modulus non-linearity in each network layer, and proved translation invariance (asymptotically in the wavelet scale parameter) and deformation stability of the corresponding feature extractor. This paper complements Mallat's results by developing a theory that encompasses general convolutional transforms, or in more technical parlance, general semi-discrete frames (including Weyl-Heisenberg filters, curvelets, shearlets, ridgelets, wavelets, and learned filters), general Lipschitz-continuous non-linearities (e.g., rectified linear units, shifted logistic sigmoids, hyperbolic tangents, and modulus functions), and general Lipschitz-continuous pooling operators emulating, e.g., sub-sampling and averaging. In addition, all of these elements can be different in different network layers. For the resulting feature extractor we prove a translation invariance result of vertical nature in the sense of the features becoming progressively more translation-invariant with increasing network depth, and we establish deformation sensitivity bounds that apply to signal classes such as, e.g., band-limited functions, cartoon functions, and Lipschitz functions.

I work with a lot of data science teams at our enterprise customers, and in the past several months, I've seen an increased adoption of machine learning and deep learning frameworks for a wide range of applications. As with other use cases in big data analytics and data science, these data science teams want to run their preferred deep learning frameworks and tools in Docker containers on the BlueData EPIC software platform. So part of my job is trying out these cool new tools and making sure they run as they should on our platform -- and to help develop new functionality that might solve any challenges. One of the most popular open-source frameworks for deep learning and machine learning is TensorFlow. TensorFlow was originally developed by researchers and engineers working at Google to conduct machine learning for deep neural networks research.