Each week at CompassRed, we take a few minutes to share what we've been reading with you.  Some of it's technical, some of it's topical, all of it's interesting. Want these fresh in your inbox every Monday? Subscribe now!


Quote of the week: "Hiding within those mounds of data is knowledge that could change the life of a patient, or change the world."  Atul Butte, Stanford
 

1. Leading with AI: We'll soon find that artificial intelligence has made it's way into every nook and cranny of our lives, augmenting things that we already do to make them better.  Naturally, that means that companies will want to see AI make its way into their sales cycles, improving performance for customers, sales people, and the bottom line.  VentureBeat reports that Afiniti, an AI-first sales boosting company, is about to splash the market as a $1.6 billion IPO.  Perhaps that's what's next.  A decade ago companies made a splash by being "mobile-first".  Now with AI being more heavily adopted, we see "AI-first" companies leading the way.
 

2. Self-Driving...Trucks?: The author of this illustrious newsletter personally looks forward to the days of not driving his car.  Conveniently, it seems that every auto-company is making it their personal mission to make that dream a reality.  What will probably come sooner, though, is an automated truck.  Elon "definitely-not-a-super-villain" Musk teased that Tesla will reveal an electric semi-truck in September.  This paves the way one step closer to full automation, which will be powered by computer vision and machine learning techniques.  We wonder - if we automate all of the trucks, what happens to all of the jobs of truck drivers?

3. Computers Draw Like 4 Year Olds:  In each of the past two weeks, we explored how well machines could see, ultimately deciding humans are still far superior to our technological counterparts.  Well, Google being Google, they decided to test how well machines could draw.  The images below show a human's attempt at drawing pigs followed by a machine's recreation of them.  You'll notice that the machine mostly plays follow the leader, attempting to fix the human "mistakes".  The 8-legged pig ended up with the correct number of legs.  The truck became a...pig truck(?).

4. Democratizing Machine Learning: Often, new technologies begin within large companies before being dispersed for use by the masses.  It is this dispersal, though, that leads us to unique growth and insights.  For years, machine learning and artificial intelligence have largely been in the hands of large corporations with the resources to make those technologies possible. As technology improves, though, this opinion piece from Wharton believes that "the democratization of ML gives individuals and startups a chance to get their ideas off the ground and prove their concepts before raising the funds needed to scale."  In fact, "just like cloud computing ushered in the current explosion in startup … machine learning platforms will likely power the next generation of consumer and business tools."
 

5. What on Earth is AI?  At CompassRed, we often talk about "artificial intelligence".  Almost as often as we talk about AI, we get asked the question, "What on Earth is AI?" Fair question.  The answer?  It's complicated.  Dr. Steve Poulin, our Chief Data Scientist, breaks AI down in our blog post.

Reading something great that we should share? Email us at Labs@compassRed.com

Data That's Beautiful

So, About Planes
Rumor has it that some kind of incident happened on a flight last week. As it turns out, some airlines have a bit more of a penchant to "involuntarily deny boarding" passengers.  We bet you can't guess which airline does this the most (spoiler: you totally can).

Data We're Watching

Addresses: Now Open
For lots of data, one of the best ways to visualize it is by mapping it.  The problem is that we need the appropriate tools to make that possible - whether those are addresses or shapes of regions (like a county or state).  Those files, particularly addresses, aren't always easy to find.  Now Open Addresses has released a dataset of 452,894,446 addresses, making it easier than ever to map projects - whatever their size.

Posted
AuthorRyan Harrington