To whom correspondence should be addressed; Email: artitw@gmail.com Customer temporal behavioral data was represented as images in order to perform churn prediction by leveraging deep learning architectures prominent in image classification. Supervised learning was performed on labeled data of over 6 million customers using deep convolutional neural networks, which achieved an AUC of 0.743 on the test dataset using no more than 12 temporal features for each customer. Unsupervised learning was conducted using autoencoders to better understand the reasons for customer churn. Images that maximally activate the hidden units of an autoencoder trained with churned customers reveal ample opportunities for action to be taken to prevent churn among strong data, no voice users.

A function $f: \mathbb{R}^d \rightarrow \mathbb{R}$ is referred to as a Sparse Additive Model (SPAM), if it is of the form $f(\mathbf{x}) = \sum_{l \in \mathcal{S}}\phi_{l}(x_l)$, where $\mathcal{S} \subset [d]$, $|\mathcal{S}| \ll d$. Assuming $\phi_l$'s and $\mathcal{S}$ to be unknown, the problem of estimating $f$ from its samples has been studied extensively. In this work, we consider a generalized SPAM, allowing for second order interaction terms. For some $\mathcal{S}_1 \subset [d], \mathcal{S}_2 \subset {[d] \choose 2}$, the function $f$ is assumed to be of the form: $$f(\mathbf{x}) = \sum_{p \in \mathcal{S}_1}\phi_{p} (x_p) + \sum_{(l,l^{\prime}) \in \mathcal{S}_2}\phi_{(l,l^{\prime})} (x_{l},x_{l^{\prime}}).$$ Assuming $\phi_{p},\phi_{(l,l^{\prime})}$, $\mathcal{S}_1$ and, $\mathcal{S}_2$ to be unknown, we provide a randomized algorithm that queries $f$ and exactly recovers $\mathcal{S}_1,\mathcal{S}_2$. Consequently, this also enables us to estimate the underlying $\phi_p, \phi_{(l,l^{\prime})}$. We derive sample complexity bounds for our scheme and also extend our analysis to include the situation where the queries are corrupted with noise -- either stochastic, or arbitrary but bounded. Lastly, we provide simulation results on synthetic data, that validate our theoretical findings.

Recent research has identified independent predictors of dementia to include age at diagnosis, transient ischemic attack and stroke status, and years of education, with vascular factors playing a greater role in disease pathogenesis than previously thought. The findings were presented at the 2016 annual meeting of the American Academy of Neurology (AAN). In the abstract, the researchers wrote that dementia encompasses a broad set of neurologic diseases, producing progressive declines in memory and/or thinking faculties, sometimes alongside personality and emotional disturbances. "Worldwide, approximately 35.6 million people have dementia, and this number is only expected to grow due to an aging population," they wrote. "Unfortunately, it is exceedingly difficult to predict who will develop dementia, let alone what type. This makes it difficult to mobilize various preventive strategies supported by mounting evidence."

In TensorFlow, during the training process of the models, the parameters are maintained in the memory as variables. When a variable is created, you can use a tensor defined as a parameter of the function as an initial value, which can be a constant or a random value.

Welcome to the 8th part of our machine learning regression tutorial within our Machine Learning with Python tutorial series. Where we left off, we had just realized that we needed to replicate some non-trivial algorithms into Python code in an attempt to calculate a best-fit line for a given dataset. Before we embark on that, why are we going to bother with all of this? Linear Regression is basically the brick to the machine learning building. It is used in almost every single major machine learning algorithm, so an understanding of it will help you to get the foundation for most major machine learning algorithms. For the enthusiastic among us, understanding linear regression and general linear algebra is the first step towards writing your own custom machine learning algorithms and branching out into the bleeding edge of machine learning, using what ever the best processing is at the time.

In this special guest feature, Dave O'Flanagan, CEO and co-founder of Boxever, outlines how airlines are leveraging big data and predictive capabilities to transform how they engage with customers. Dave is the CEO and co-founder of Boxever, a data science and omni-channel personalization platform for travel companies. The subprime financial crisis revealed that our data is only as good as our ability to analyze and understand it. AI will be necessary to helping prevent the next crisis before it happens. Marco Scirea, a PhD student at the IT University of Copenhagen, won the best paper award at the EvoMUSART conference for his research on music composition using artificial intelligence.

A computer can probably beat you at chess and no one goes anywhere without a GPS. Transhumanist prophet Ray Kurzweil says we will ascend into being computers in a few years (though he also claims solar power will out-produce fossil fuels in a decade, so use caution when he is selling books) but some think it's the other way around, and that humans will instead be the ultimate supercomputers. Danish physicist Jacob Sherson, writing about his beliefs in Nature, said, "It may sound dramatic, but we are currently in a race with technology -- and steadily being overtaken in many areas. Features that used to be uniquely human are fully captured by contemporary algorithms. Our results are here to demonstrate that there is still a difference between the abilities of a man and a machine."

IBM announced a deal with sports and fitness retailer Under Armour Inc. to use machine learning technology from Watson and showed off an application for diabetic care developed with the supercomputer's data, highlighting the company's effort to expand Watson's capabilities for the health-care industry. IBM and Under Armour released an updated fitness application for Apple Inc.'s iPhones that uses data powered by Watson, IBM Chief Executive Officer Ginni Rometty said Wednesday in a speech at the International Consumer Electronics Show in Las Vegas. Separately, Medtronic Plc CEO Omar Ishrak joined Rometty on stage to unveil a prototype for a diabetes-management app that tests have shown may be capable of predicting hypoglycemic events as early as three hours in advance. The application still needs to go through regulatory review -- it will roll out this summer, Rometty said. The ability to predict the hypoglycemic events is a "breakthrough," she said.

Nasa has announced that it has found evidence of flowing water on Mars. Scientists have long speculated that Recurring Slope Lineae -- or dark patches -- on Mars were made up of briny water but the new findings prove that those patches are caused by liquid water, which it has established by finding hydrated salts. Several hundred camped outside the London store in Covent Garden. The 6s will have new features like a vastly improved camera and a pressure-sensitive "3D Touch" display

COVER Intelligence is hard to define, but you know it when you see it … Or do you? Artificial intelligence researchers can now design algorithms with almost humanlike abilities to perceive images, communicate with language, and learn from experience. Can we learn anything about how our neuron-based minds work from these machines? Do we need to worry about what these algorithmic minds might be learning about us? On the cover is a visualization of human brain connectivity from MRI diffusion imaging, with superimposed computer connectors.