Data School's 8-week Master Course begins September 28. More than two-thirds of the available spots are gone! Learn more about the course and enroll: http://www.dataschool.io/learn/ This info session was recorded on September 13. View the chat history and complete Q&A: http://ccst.io/e/text-course

Secret to GoogleBrain and RankBrain algorithm revealed. One is going to give a historical overview about GoogleBrain and analyse the pattern, then we will conclude our finding about the current situation and future changes in search engine algorithm. Back in 2006 there were some interests in implementing artificial intelligence in Google search engine algorithm. A few years later in 2014, GoogleBrain was established after acquisition of DeepMind, a British artificial intelligence company which was founded in 2010. They worked on how to play video games based on machine learning and artificial neural networks (ANNs).

The next paper from our reading group is by Song Han, Huizi Mao, and William J. Dally. It won the best paper award at ICLR 2016. It details three methods of compressing a neural network in order to reduce the size of the network on disk, improve performance, and decrease run time. Pre-trained convolutional neural networks are too large for mobile devices: AlexNet is 240 MB and VGG-16 is over 552 MB. This seems small when compared to a music library or large video, but the difference is that the networks reside in memory when running.

Attention everyone who's smugly proclaimed that they "just don't need a full night's sleep." You might have fooled us coffee-chuggers before, but now there's evidence that you're not quite superhuman. According to a new paper published in the journal Brain and Behavior (via Medical Xpress), University of Utah researchers studied patterns in the 839 people, dividing them into two groups: those who slept six hours or fewer per night, and those who got more. They then divided the short sleepers into two more groups: those who felt fine during the day, and those who reported feeling drowsy. When they put them in the MRI scanner -- a dark, boring tube of white noise (perfect for a little nap) -- both sets of short sleepers showed signs of sleep in their brain patterns while getting scanned.

Ever want to escape the earth? There are billions of other worlds out there, and while we don't know a lot about most of them, there's a good chance at least one is better than Earth itself. With the possible exception of Mars, it's extremely unlikely humans living now will ever make it to any of those distant worlds. What might, instead, is a Von Neumann self-replicating robot probe. Von Neumann probes are not a new idea.

Many important applications, including signal reconstruction, parameter estimation, and signal processing in a compressed domain, rely on a low-dimensional representation of the dataset that preserves {\em all} pairwise distances between the data points and leverages the inherent geometric structure that is typically present. Recently Hedge, Sankaranarayanan, Yin and Baraniuk \cite{hedge2015} proposed the first data-aware near-isometric linear embedding which achieves the best of both worlds. However, their method NuMax does not scale to large-scale datasets. Our main contribution is a simple, data-aware, near-isometric linear dimensionality reduction method which significantly outperforms a state-of-the-art method \cite{hedge2015} with respect to scalability while achieving high quality near-isometries. Furthermore, our method comes with strong worst-case theoretical guarantees that allow us to guarantee the quality of the obtained near-isometry. We verify experimentally the efficiency of our method on numerous real-world datasets, where we find that our method ($<$10 secs) is more than 3\,000$\times$ faster than the state-of-the-art method \cite{hedge2015} ($>$9 hours) on medium scale datasets with 60\,000 data points in 784 dimensions. Finally, we use our method as a preprocessing step to increase the computational efficiency of a classification application and for speeding up approximate nearest neighbor queries.

Symmetric Nonnegative Matrix Factorization (SNMF) models arise naturally as simple reformulations of many standard clustering algorithms including the popular spectral clustering method. Recent work has demonstrated that an elementary instance of SNMF provides superior clustering quality compared to many classic clustering algorithms on a variety of synthetic and real world data sets. In this work, we present novel reformulations of this instance of SNMF based on the notion of variable splitting and produce two fast and effective algorithms for its optimization using i) the provably convergent Accelerated Proximal Gradient (APG) procedure and ii) a heuristic version of the Alternating Direction Method of Multipliers (ADMM) framework. Our two algorithms present an interesting tradeoff between computational speed and mathematical convergence guarantee: while the former method is provably convergent it is considerably slower than the latter approach, for which we also provide significant but less stringent mathematical proof regarding its convergence. Through extensive experiments we show not only that the efficacy of these approaches is equal to that of the state of the art SNMF algorithm, but also that the latter of our algorithms is extremely fast being one to two orders of magnitude faster in terms of total computation time than the state of the art approach, outperforming even spectral clustering in terms of computation time on large data sets.

Image denoising is an important pre-processing step in medical image analysis. Different algorithms have been proposed in past three decades with varying denoising performances. More recently, having outperformed all conventional methods, deep learning based models have shown a great promise. These methods are however limited for requirement of large training sample size and high computational costs. In this paper we show that using small sample size, denoising autoencoders constructed using convolutional layers can be used for efficient denoising of medical images. Heterogeneous images can be combined to boost sample size for increased denoising performance. Simplest of networks can reconstruct images with corruption levels so high that noise and signal are not differentiable to human eye.

A car is about to hit a dozen pedestrians. Is it better for the car to veer off the road and kill the driver but save the pedestrians? Or is it better to save the driver and kill all those other people? That's the thorny philosophical question that the makers of autonomous vehicles -- self-driving cars -- are grappling with these days, and a new study sheds some light on what people actually want that car to do. It turns out that the answer depends on whether you are the driver of the car or not.

Four people stand, eyes squinting, on the verdant, freshly mowed field of Boston's Fenway Park. The smiles on their faces are broad, if a bit shy. They are among the lucky ones--the tiny share of people who fought and survived pancreatic cancer. The scene is from a photograph pinned to a cubicle at Beth Israel Deaconess Medical Center in Boston. "What is it about these four people on the field?" asks A. James Moser, co-director of the Pancreas and Liver Institute and director of the Pancreatic Cancer Research Institute at Beth Israel Deaconess Medical Center.