Data Science newsletter – February 1, 2017

GROUP CURATION: N/A

In our recent research at MIT Sloan Management Review, we found that only 34% of the respondents felt that they needed to improve their IoT data security. If you are trying to decide if the glass is full or empty, that glass seems two-thirds empty to me.

The research included responses from 1,480 executives, managers, and IT professionals working in a wide variety of industries. It focused on the perspective of organizations, not security professionals, and tried to understand their challenges and opportunities associated with the Internet of Things.

Hungry penguins have inspired a novel way of making sure computer code in smart cars does not crash.

On his personal blog, Geoff Boeing, an urban planning PhD candidate at the University of California, Berkeley, has created a series of illustrations that allows easy comparison of the urban forms of cities as disparate as Rome and Irvine.

Inspired by Great Streets by Allan Jacobs, Boeing used Python and the OSMnx package he created for his dissertation. “With OSMnx we can download a street network from OpenStreetMap for anywhere in the world in just one line of code,” explains Boeing.

Walmart – the world’s biggest retailer with over 20,000 stores in 28 countries, is in the process of building the world’ biggest private cloud, to process 2.5 petabytes of data every hour.

To make sense of all of this information, and put it to work solving problems, the company has created what it calls its Data Café – a state-of-the-art analytics hub located within its Bentonville, Arkansas headquarters.

Here, over 200 streams of internal and external data, including 40 petabytes of recent transactional data, can be modelled, manipulated and visualized. Teams from any part of the business are invited to bring their problems to the analytics experts and then see a solution appear before their eyes on the nerve centre’s touch screen “smart boards”.

f you had a swanky, expensive car, you might not be eager to hand off the keys to a total stranger looking to open it up or slow roll their high school reunion. You might be protective. You might be reluctant. And you might be convinced to change your mind. That is the premise at the core of Turo, a car-sharing startup with a presence in more than 4,500 cities that specializes in dope rides. Turo is building a community of car enthusiasts psyched to let other people take the wheel — if only because the money helps justify having a sweet whip.

Using a statistical method initially developed by Google, a Yale School of Public Health-led research team has devised a novel way to better analyze the impact of vaccines. The research was published in the Proceedings of the National Academy of Sciences. … A team led by Daniel Weinberger, assistant professor in the Department of Epidemiology of Microbial Diseases, used a method called “synthetic controls,” which was not previously applied in an epidemiology context, to analyze the impact of the pneumococcal vaccine.

In their recent report, “Unleashing Technology, Media, and Telecom with Digital Transformation”, the Boston Consulting Group shines a light on three trends in particular: artificial intelligence (AI), the Internet of Things (IoT), and cybersecurity.

Understanding these trends may give companies in the sector a competitive advantage. AI can help companies automate their customer support or network operations. Internet-connected devices can help technology vendors improve business processes and increase efficiencies in manufacturing, transportation, logistics, and utilities companies. And a new focus on cybersecurity can help TMT companies prevent breaches, to which they are particularly vulnerable. While AI, IoT and cybersecurity have not actively fueled growth for TMT companies in the past, they are all now active targets of venture funding and will undoubtedly shape future performance.

A new strategy promises to tame the trolls and reinvigorate civil discussion on the Internet. Hatched by Jigsaw, an in-house think tank at Google’s parent company, Alphabet (googl, -0.75%), the tool relies on artificial intelligence and could solve the once-impossible task of vetting floods of online comments.

To explain what Jigsaw is up against, chief research scientist Lucas Dixon compares the troll problem to so-called denial-of-service attacks in which attackers flood a website with garbage traffic in order to knock it off-line.

“Instead of flooding your website with traffic, it’s flooding the comment section or your social media or hashtag so that no one else can have a word, and basically control the conversation,” says Dixon.

“The striking results overturn my own theory that people with complete locked-in syndrome are not capable of communication,” said Niels Birbaumer, a neuroscientist at Switzerland’s Wyss Center for Bio and Neuroengineering, who co-led the study.

The trial, published in the journal PLOS Biology on Tuesday, involved four patients with amyotrophic lateral sclerosis (ALS) – a progressive motor neuron disease that destroys the part of the nervous system responsible for movement.

In 2000, only 8 percent of U.S. hospitals had some version of a game-changing computer technology called the Picture Archiving and Communications System, or PACS. By 2008, more than three out of four did.

Because radiology was the first medical specialty to computerize, what has happened to it — at once shocking and, in retrospect, entirely predictable — is our canary in the digital coal mine, its experience offering important lessons for patients, clinicians and health care systems.

Another game just fell to the machines.

Yesterday, after 20 days of play at a casino in Pittsburgh, an AI built by two Carnegie Mellon researchers officially defeated four top players at no-limit Texas Hold ‘Em—a particularly complex form of poker that relies heavily on longterm betting strategies and game theory. Over the past twenty years, machines have topped the best humans at checkers, chess, Scrabble, Jeopardy!, and even the ancient game of Go. But no AI had ever beaten the best at such an extreme game of “imperfect information,” a game where certain elements, such as the cards on the table, are hidden. Among humans, no-limit Hold ‘Em requires a certain degree of intuition, not to mention luck.

You believe that automated machine learning is going to transform data science. What major changes do you expect to see in your lifetime?

I can easily see automated machine learning transforming the field of data science within the next 5 years. In essence, most advancements in data science, AI, and machine learning lately— for example, AlphaGo dominating the world’s best Go players—have been made by large teams of researchers and engineers meticulously designing and optimizing complex machine learning systems that specialize in one particular problem. Automated machine learning seeks to augment and perhaps one day replace those teams of researchers with AI systems that follow the same design and optimization process, but aren’t constrained by human biases. As an added bonus, automated machine learning systems won’t be limited by human needs such as food and sleep, so they can work as long as there’s computer hardware and electricity.

For artificial general intelligence (AGI) it would be efficient if multiple users trained the same giant neural network, permitting parameter reuse, without catastrophic forgetting. PathNet is a first step in this direction. It is a neural network algorithm that uses agents embedded in the neural network whose task is to discover which parts of the network to re-use for new tasks. Agents are pathways (views) through the network which determine the subset of parameters that are used and updated by the forwards and backwards passes of the backpropogation algorithm. During learning, a tournament selection genetic algorithm is used to select pathways through the neural network for replication and mutation. Pathway fitness is the performance of that pathway measured according to a cost function. We demonstrate successful transfer learning; fixing the parameters along a path learned on task A and re-evolving a new population of paths for task B, allows task B to be learned faster than it could be learned from scratch or after fine-tuning. Paths evolved on task B re-use parts of the optimal path evolved on task A. Positive transfer was demonstrated for binary MNIST, CIFAR, and SVHN supervised learning classification tasks, and a set of Atari and Labyrinth reinforcement learning tasks, suggesting PathNets have general applicability for neural network training. Finally, PathNet also significantly improves the robustness to hyperparameter choices of a parallel asynchronous reinforcement learning algorithm (A3C).

The Julia language provides a fresh new approach to numerical computing, where there is no longer a compromise between performance and productivity. A high-level language that makes writing natural mathematical code easy, with runtime speeds approaching raw C, Julia has been used to model economic systems at the Federal Reserve, drive autonomous cars at University of California Berkeley, optimize the power grid, calculate solvency requirements for large insurance firms, model the US mortgage markets and map all the stars in the sky.

It would be no surprise then that Julia is a natural fit in many areas of machine learning. ML, and in particular deep learning, drives some of the most demanding numerical computing applications in use today. And the powers of Julia make it a perfect language to implement these algorithms.

We introduce a new algorithm named WGAN, an alternative to traditional GAN training. In this new model, we show that we can improve the stability of learning, get rid of problems like mode collapse, and provide meaningful learning curves useful for debugging and hyperparameter searches. Furthermore, we show that the corresponding optimization problem is sound, and provide extensive theoretical work highlighting the deep connections to other distances between distributions.

Modern advanced analytics applications make use of machine learning techniques and contain multiple steps of domain-specific and general-purpose processing with high resource requirements. We present KeystoneML, a system that captures and optimizes the end-to-end large-scale machine learning applications for high-throughput training in a distributed environment with a high-level API. This approach offers increased ease of use and higher performance over existing systems for large scale learning. We demonstrate the effectiveness of KeystoneML in achieving high quality statistical accuracy and scalable training using real world datasets in several domains.

Traditionally the experience of a digital service follows pre-defined user journeys with clear states and actions. Until recently, it has been the designer’s job to create these linear workflows and transform them into understandable and unobtrusive experiences. This is the story of how that practice is about to change.

This guide will step you through configuring an Amazon Alexa Skill and Lambda function to control the relay attached to an IoT device — specifically, an Intel Edison-powered starter kit. But with some simple modification to the code, you could make Alexa control nearly anything you like.

After working on a few data science (aka data analytics aka machine learning) problems with geoscientific data, I think we’ve figured out the 10-step workflow. I’m happy to share it with you now:

Look at all these cool problems, machine learning can solve all of these! I just need to figure out which model to use, parameterize it, and IT’S GONNA BE AWESOME, WE’LL BE RICH. Let’s just have a quick look at the data…

Oh, there’s no data.