Revolutions

New round of R Consortium grants announced

The R Consortium has just announced its latest round of project grants. After reviewing the proposals submitted by the R community, the Infrastructure Steering Committee has elected to fund the following projects for the Spring 2018 season:

• Further updates to the DBI package, to provide a consistent interface between R and databases
• Updating the infrastructure in R for building binary packages on Windows and Mac
• A collaboration with Statisticians in the Pharmaceutical Industry to validate R packages to regulatory standards
• A project to validate and maintain some of the historic numerical algorithms used by R 
• A working group to facilitate working with US census data in R
• Consolidation of tools and workflows for handling missing data in R
• Templates and tools for the development of teaching resources with R

This latest round of grants brings the total funding for projects proposed by R community members to more than $650,000. The R Consortium itself is funded by its members (including my own employer, Microsoft), so if you'd like to see more such projects why not ask your own employer to become a member? 

R Consortium: Announcing the R Consortium ISC Funded Project grant recipients for Spring 2018

 

Because it's Friday: Bad road

Sometimes I think the potholes in the roads in Chicago are bad, but then a road like this puts things into perspective:

(Thanks to TH for the link.) Don't miss the shots looking back near the end to see how many people are in that vehicle!

That's all from the blog for this week. Have a great weekend, and we'll be back after the US holiday on Monday. Enjoy!

Reflections on the ROpenSci Unconference

I had an amazing time this week participating in the 2018 ROpenSci Unconference, the sixth annual ROpenSci hackathon bringing together people to advance the tools and community for scientific computing with R. It was so inspiring to be among such a talented and dedicated group of people — special kudos goes to the organizing committee for curating such a great crowd. (I heard there were over 200 hundred nominations from which the 65 or so attendees were selected.)

The idea behind the unconference is to spend two full days hacking on projects of interest to the community. Before the conference begins, the participants suggest projects as Github issues and begin discussions there. On the first day of the conference (after an icebreaker), the participants vote for projects they'd be interested in working on, and then form up into groups of 2-6 people or so to work on them. And then everyone gets to work! You can get a sense of the activity by looking at the #runconf18 hashtag on Twitter (and especially the photos).

I joined the "Tensorflow Probability for R" team, where we worked mainly with the greta package, which uses Tensorflow Probability to implement a Markov-Chain Monte-Carlo system to find solutions to complex statistical models. I hadn't used greta before, so I focused on trying out some simple examples to understand how greta works. In the process I learned that greta is a really powerful package: it solves many of the same problems as stan, but with a really elegant R interface that generalizes beyond Bayesian models. (Over the next couple of weeks I'll elaborate the examples for the greta package and write a more detailed blog post.)

At the end of the second day, everyone gets together to "report out" on their projects: a three minute presentation to review the progress from the hackathon. You can browse the list of projects here: follow the links to the Github repositories and check out the Readme.md files for details on each project.

A sincere thank you to all participants in #runconf18

This thread👇includes links to all project repos: https://t.co/2PhAz4zSuK#rstats pic.twitter.com/8SICcWkQ0v

— rOpenSci (@rOpenSci) May 25, 2018

On a personal note, it was also a great joy that my team could sponsor the unconference and provide the venue for this year's event. The Microsoft Reactor in Seattle turned out to be a great place to hold a hackathon like this, with plenty of space for everyone to form into small teams and find a comfortable place to work. This kind of event is exactly why we are opening Reactor spaces around the world, as spaces for the community to meet, gather and work, and it was great to see that vision realized in such a great way this week. 

Thanks once again to all of the participants and especially the organizers (with a special shout-out to Stefanie Butland) for such a wonderful event. I can't wait for the next one!

AI, Machine Learning and Data Science Roundup: May 2018

A monthly roundup of news about Artificial Intelligence, Machine Learning and Data Science. This is an eclectic collection of interesting blog posts, software announcements and data applications I've noted over the past month or so.

Open Source AI, ML & Data Science News

R 3.5.0 has been released, with significant performance improvements.

Project Jupyter has received the ACM Software System Award.

PyTorch 1.0 has been released, and deemed ready for both research and production for AI applications.

The ecosystem for ONNX (the open standard for exchange of neural network models) expands, with official support for Core ML, NVIDIA TensorRT 4, and the Snapdragon Neural Processing Engine.

A new Swift API for Tensorflow has been released, with compiler and language enhancements.

TensorFlow Probability, a probabilistic programming toolbox for machine learning.

Industry News

Impressive benchmarks for training Imagenet and CIFAR10 using AWS spot instances, with Pytorch and fastai.

Oracle acquires machine learning platform Datascience.com, a cloud workspace platform for data science projects and workloads.

Microsoft News

Microsoft Build featured several announcements regarding AI and machine learning. You can watch the keynotes here.

Many new capabilities and features for Cognitive Services were announced at BUILD, including: Bing Custom Search, Custom Vision Service and Custom Decision Service; Microsoft’s Cognitive Services Labs (with previews of emerging Cognitive Services technologies); and Video Indexer.

Cognitive Search, a new Azure service providing an AI-first approach to content understanding.

New Python packages for machine learning in Azure: Forecasting, Computer Vision, and Text Analytics.

ML.net is a new open-source, cross-platform machine learning framework for developers. The technology behind AI features in Office and Windows has now been released as a project on Github.

Project Brainwave is now in preview, bringing hardware-accelerated inference for AI to Azure. Deep neural networks encoded by FPGAs (field-programmable gate arrays) can dramatically reduce the time to classify images (as just one application example). ResNet50 is available in Project Brainwave now (subject to quota approval).

Microsoft and Qualcomm have partnered on the Vision AI Developer Kit, now in preview.

Updates to the Bot Builder SDK and Bot Framework Emulator.

Microsoft R Open 3.4.4 is now available.

Learning resources

Transfer Learning for Text using Deep Learning Virtual Machine (DLVM). Code comparing the performance of eight machine-comprehension algorithms is also available in Github.

The TWIML AI podcast series on differential privacy, a technique for collecting data in such a way that it reduces the impact on privacy even in the event of a leak.

Building Neural Networks with TensorFlow and the Azure Data Science Virtual Machine. This hands-on lab walks through the process of building an image recognizer using transfer learning with MobilenetV1.

Full-integrated experience simplifying Language Understanding in conversational AI systems.

Deep Learning Image Segmentation for Ecommerce Catalogue Visual Search, with details on removing the background from a product image using GrabCut and Tiramisu.

How to Develop a Currency Detection Model using Azure Machine Learning, with details on how the real-time banknote recognition capability of the Seeing AI application was implemented in CoreML.

Deep Learning for Emojis with VS Code Tools for AI: semantic analysis of text with emojis.

A maze-solving Minecraft robot, implemented in R.

Find previous editions of the monthly AI roundup here.

Video: speeding up R with parallel programming in the cloud

I had a great time in Budapest last week for the eRum 2018 conference. The organizers have already made all of the videos available online. Here's my presentation: Speeding up R with Parallel Programming in the cloud.

You can find (and download) my presentation slides here. And if you just want the references from the last slide, here are the links:

• Find foreach on CRAN: CRAN.R-project.org/package=foreach

• Code for birthday problem simulation: cda.ms/7d

• Docs for the foreach package: cda.ms/tc

• Get doAzureParallel: cda.ms/7w

• Get Azure subscription with $200 free credits: cda.ms/td

Because it's Friday: Laurel or Yanny

I can only assume you've heard about this already: it's gone wildly viral in the USA, and I assume elsewhere in the world. But it's a lovely example of an auditory illusion, and as regular readers know I like to collect such things (like this one and this one).

What do you hear?! Yanny or Laurel pic.twitter.com/jvHhCbMc8I

— Cloe Feldman (@CloeCouture) May 15, 2018

Kottke has a useful roundup of the best scientific coverage of why some people hear "laurel" and others hear "yanny". Linguist Rachel Gutman also provides a detailed analysis for The Atlantic. For the record though, all I can hear is "laurel", and that indeed is what the actor said for a recording for Vocabulary.com. 

That's all from us for this week. Next week I'll be at the ROpenSci unconference in Seattle, but we'll back with more on the blog on Monday. In the meantime have a great weekend!

Let me tell you what you missed at BUILD

If you weren't able to attend last weeks' BUILD conference in Seattle, you can always catch up on the keynotes and the session talks online, or read this recap by Charlotte Yarkoni. Or, if you have 45 minutes on Wednesday next week you can join Tim Heuer and me as we recap some of the biggest announcements of the conference, and few things you might have missed as well.  We'll focus mainly on the AI and Machine Learning related announcements, including:

• Project Brainwave, and the performance you can get from hardware-accelerated, real-time AI
• New, platform-agnostic options for AI models, including ONNX and docker containers
• New Azure ML Python packages for forecasting, computer vision, and text analytics
• Projects from the Microsoft AI Lab and AI for Accessibility
• ... and a few extra surprises as well.

The webinar will be at 10AM Pacific time on Wednesday, May 23 when you'll also have the opportunity to ask us live questions. You can register for the webinar below, which will also give you access to the on-demand recording if you can't make the live sessions. We hope to see you there!

Azure Webinar Series: Top Azure Takeaways from Microsoft Build

 

Mind Bytes: Solving Societal Challenges with Artificial Intelligence

By Francesca Lazzeri (@frlazzeri), Data Scientist at Microsoft

Artificial​ ​intelligence​ ​(AI)​ solutions​ ​are playing a growing role in our everyday life, ​​and​ ​are​ ​being adopted​ ​broadly, in private and public domains. ​ ​While​ ​the​ ​notion​ ​of​ ​AI​ ​has​ ​been around​ ​for​ ​over​ ​sixty​ ​years, real-world​ ​AI scenarios and applications​ ​have​ ​only​ ​increased​ ​in​ ​the​ ​last​ ​decade​ ​due​ ​to​ ​three​ ​simultaneous events: ​ ​​improved​ ​computing​ ​power, capability​ ​to​ ​capture​ ​and​ ​store​ ​massive​ ​amounts​ ​of​ ​​​data, and faster​ ​algorithms.  

AI solutions help determine the commercials you see online, the movie you will watch with your family, the routes you may take to get to work. Beyond​ ​the​ most popular apps, ​these​ ​systems​ ​are​ ​also​ ​being​ ​implemented​ ​in​ ​critical areas​ such as​ health care, immigration policy, ​finance, ​and​ the ​workplace. The design​ ​and​ ​implementation​ ​of​ ​these AI ​tools​ ​presents​ ​deep societal​ ​challenges​ ​​that​ ​will​ ​shape​ ​our​ ​present​ ​and near​ ​future. ​ ​ ​

To identify and contribute to the current dialog around the emerging societal challenges that AI is bringing, we attended Mind Bytes at the University of Chicago. Mind Bytes is an annual research computing symposium and exposition designed to showcase to more than 200 attendees cutting-edge research and applications in the field of AI. Some of the most interesting demos and posters presented were:

• An Online GIS Platform to Support China’s National Park System establishment – Manyi Wang
• 17 Years in Chicago Neighborhoods: Mapping Crime Trajectories in 801 Census Tracts – Liang CAI
• Characterizing the Ultrastructural Determinants of Biophotonic Reflectivity in Cephalopod Skin: A Challenge for 3D Segmentation – Stephen Senft, Teodora Szasz, Hakizumwami B. Runesha, Roger T. Hanlon
• Exploring Spatial Distribution of Risk Factors for Teen Pregnancy – Emily Orenstein and Iris Mire

The Mind Bytes panel on Solving Societal Challenges with Artificial Intelligence represented an incredible occasion for us to interact with students and many other AI experts from the field, and understand how we can work together to ensure that AI is developed in a responsible manner, so that people will trust it and deploy it broadly, both to increase business and personal productivity and to help solve societal problems.

Specifically, the panel focused on three fundamental areas of discussion related to current and future AI aspects:

• What areas do you see AI most successfully applied?
• What is the major challenge that you think should be met before getting the full benefit of AI?
• What can researchers and students do now to build a system able to address those challenges?

The following sections aim at answering these questions in more detail and reflect on the latest academic and industry research. AI is already with us, and we are now faced with important choices on how it will be designed and applied. Most promisingly, the approaches observed at Mind Bytes demonstrate that there is growing interest in developing AI that is attuned to underlying issues of fairness and equality.

Because it's Friday: Planes, pains, and automobiles

It's conference season, so I've been travelling on a lot of planes recently. (And today, I'm heading to Budapest for eRum 2018.) So this resonates with me right now:

There's another video in the same vein about hotels which is quite funny as well. Anyway, that's all from us for this week; we'll be back next week reporting from eRum. Until then, have a great weekend!

In case you missed it: April 2018 roundup

In case you missed them, here are some articles from April of particular interest to R users.

Microsoft R Open 3.4.4, based on R 3.4.4, is now available.

An R script by Ryan Timpe converts a photo into instructions for rendering it as LEGO bricks.

R functions to build a random maze in Minecraft, and have your avatar solve the maze automatically.

A dive into some of the internal changes bringing performance improvements to the new R 3.5.0.

AI, Machine Learning and Data Science Roundup, April 2018. 

An analysis with R shows that Uber has overtaken taxis for trips in New York City.

News from the R Consortium: new projects, results from a survey on licenses, and R-Ladies is promoted to a top-level project.

A talk, aimed at Artificial Intelligence developers, making the case for using R.

Bob Rudis analyzes data from the R-bloggers.com website, and lists the top 10 authors.

An R-based implementation of Silicon Valley's "Not Hotdog" application.

An R package for creating interesting designs with generative algorithms.

And some general interest stories (not necessarily related to R):

• Learn regular expressions with interactive "crosswords"
• A tongue-in-cheek review of Wes Anderson's movies
• Snarky captions for incorrect maps used in news reports
• A few podcast recommendations, from me

As always, thanks for the comments and please send any suggestions to me at [email protected]. Don't forget you can follow the blog using an RSS reader, via email using blogtrottr, or by following me on Twitter (I'm @revodavid). You can find roundups of previous months here.