Machine Learning :: Human Learning

AI - “artificial intelligence” - was introduced at a science conference at Dartmouth University in 1956. Back then it was a theory, but in the past few decade it has become something beyond theoretical. been less theory and more in practice than decades before.

The role of AI in education is still more theory than practice.

A goal in AI is to get machines to learn. I hesitate to say "think" but that is certainly a goal too. I am reading The Innovators: How a Group of Hackers, Geniuses, and Geeks Created the Digital Revolution currently and in that history there is a lot of discussion of the people trying to get machines to do more than just compute (calculate) but to learn from its experiences without requiring a human to program those changes. The classic example is the chess playing computer that gets better every time it wins or loses. Is that "learning?"

But has it had an impact on how you teach or how your students learn?

It may have been a mistake in the early days of AI and computers that we viewed the machine as being like the human brain. It is - and it isn't.

But neuroscientists are now finding that they can also discover more about human learning as a result of machine learning. An article on points to several interesting insights from the machine and human learning research that may play a role in AI in education.

One thing that became clear is that physical environment is something humans learn easier than machines. After a child has started walking or opened a few doors or drawers or climbed a few stairs, she learns how to do it. Show her a different door, drawer, or a spiral staircase and it doesn't make much of a difference. A robot equipped with some AI will have a much steeper learning curve to learn these simple things. It also has a poor sense of its "body." Just watch any videos online of humanoid robots trying to do those things and you'll see how difficult it is for a machine.

Then again, it takes a lot longer for humans to learn how to drive a car on a highway safely. And even when it is learned, our attention, or lack thereof, is a huge problem. AI in vehicles is learning how to drive fairly rapidly, and its attention is superior to human attention. Currently, it is still a fall back human error in mist cases, but that will certainly change in a decade or two. I learned to parallel park a car many years ago and I am still lousy at doing it. A car can do it better than me.

Although computers can do tasks they are programmed to do without any learning curve, for AI to work they need to learn by doing - much like humans. The article points out that AI systems that traced letters with robotic arms had an easier time recognizing diverse styles of handwriting and letters than visual-only systems. 

AI means a machine gets better at a task the more it does it, and it can also apply that learning to similar but not identical situations. You can program a computer to play notes and play a series of notes as a song, but getting it to compose real music requires AI.

Humans also learn from shared experiences. A lot of the learning in a classroom comes from interactions between the teacher and students and student to student. This makes me feel pretty confident in the continued need for teachers in the learning process.

One day, I am sure that machines will communicate with each other and learn from each other. This may be part of the reason that some tech and learning luminaries like Elon Musk have fears about AI

I would prefer my smart or autonomous vehicle to "talk" to other vehicles on the roads nearby and share information on traffic, obstructions and vehicles nearby with those quirky human drivers only.

AI built into learning systems, such as an online course, could guide the learning path and even anticipate problems and offer corrections to avoid them. Is that an AI "teacher" or the often-promoted "guide on the side?"

This year on the TV show Humans, one of the human couples goes for marriage counseling with a "synth" (robot). She may be a forerunner of a synth teacher.

Humans TV
The counselor (back to us) can read the husband's body language and knows he does not like talking to a synth marriage counselor.


Is Education Ready to Connect to the Internet of Things?


I first encountered the term "Internet of Things" (IoT) in 2013. It is the idea that "things" (physical devices) would be connected in their own network(s). The talk was that things in your home, office and vehicles would be wirelessly connected because they were embedded with electronics, software, sensors, actuators, and network connectivity. Things would talk to things. Things would collect and exchange data.

Some of the early predictions seemed rather silly. Taking a tagged carton of milk out of the refrigerator and not putting it back would tell my food ordering device (such as an Amazon Echo) that I was out of milk. My empty Bluetooth coffee mug would tell the Keurig coffeemaker to make me another cup.

But the "smart home" - something that pre-dates the Internet - where the HVAC knew I was almost home and adjusted the temperature off the economical setting to my comfort zone and maybe put on the front light and started dinner, was rather appealing.

In 2014, the EDUCAUSE Learning Initiative (ELI) published its “7 Things You Should Know About the Internet of Things. The Internet of Things (and its annoying abbreviation of IoT) sounded rather ominous as I imagined them proliferating across our social and physical landscapes. The ELI report said “the IoT has its roots in industrial production, where machine-to-machine communication enabled the manufacture of complex items, but it is now expanding in the commercial realm, where small monitoring devices allow such things as ovens, cars, garage doors, and the human heartbeat to be checked from a computing device.”

Some of the discussions have also been about considerations of values, ethics and ideology, especially if you consider the sharing of the data gathered. 

As your watch gathers data about your activity, food intake and heart rate, it has valuable data about your health. I do this on my Fitbit with its app. Perhaps you share that with an online service (as with the Apple watch & Apple itself) in order to get further feedback information about your health and fitness and even recommendations about things to do to improve it. If you want a really complete analysis, you are asked (hopefully) to share your medications, health history etc. Now, what if that is shared with your medical insurer and your employer?

Might we end up with a Minority Report of predictive analytics that tell the insurance company and your employer whether or not you are a risk?

Okay, I made a leap there, but not a huge one. 

This summer, EDUCAUSE published a few articles on IoT concerning higher education and the collaboration required for the IoT to work. I don't see education at any level really making significant use of IoT right now, though colleges are certainly gathering more and more data about students. That data might be used to improve admissions. Perhaps, your LMS gathers data about student activity and inactivity and can use it to predict what students need academic interventions.

It's more of an academic challenge to find things that can be used currently.

History Lesson: Way back in 1988, Mark Weiser talked about computers embedded into everyday objects and called this third wave "ubiquitous computing." Pre-Internet, this was the idea of many computers, not just the one on your desk, for one person. Add ten years and in 1999, Keven Ashton posited a fourth wave which he called the Internet of Things.

Connection was the key to both ideas. It took another decade until cheaper and smaller processors and chipsets, growing coverage of broadband networks, Bluetooth and smartphones made some of the promises of IoT seem reasonable. 

Almost any thing could be connected to the Internet. We would have guessed at computers of all sizes, cars and appliances. I don't think things such as light bulbs would have been on anyone's list.

Some forecasters predict 20 billion devices will be connected by 2020; others put the number closer to 40-100+ billion connected devices by that time.

And what will educators do with this?

Cognizant Computing in Your Pocket (or on your wrist)

Two years ago, I wrote about the prediction that your ever-smarter phone will be smarter than you by 2017. We are half way there and I still feel superior to my phone - though I admit that it remembers things that I can't seem to retain, like my appointments, phone numbers, birthdays and such.

The image I used on that post was a watch/phone from The Jetsons TV show which today might make you think of the Apple watch which is connected to that ever smarter phone.

But the idea of cognizant computing is more about a device having knowledge of or being aware of your personal experiences and using that in its calculations. Smartphones will soon be able to predict a consumer’s next move, their next purchase or interpret actions based on what it knows, according to Gartner, Inc.

This insight will be performed based on an individual’s data gathered using cognizant computing — "the next step in personal cloud computing.

“Smartphones are becoming smarter, and will be smarter than you by 2017,” said Carolina Milanesi, Research Vice President at Gartner. “If there is heavy traffic, it will wake you up early for a meeting with your boss, or simply send an apology if it is a meeting with your colleague."

The device will gather contextual information from your calendar, its sensors, your location and all the personal data  you allow it to gather. You may not even be aware of some of that data it is gathering. And that's what scares some people.

watchWhen your phone became less important for making phone calls and added apps, a camera, locations and sensors, the lines between utility, social, knowledge, entertainment and productivity got very blurry.

But does it have anything to do with learning?

Researchers at Pennsylvania State University already announced plans to test out the usefulness in the classroom of eight Apple Watches this summer.

Back in the 1980s, there was much talk about Artificial Intelligence (AI). Researchers were going to figure out how we (well, really how "experts") do what they do and reduce those tasks to a set of rules that a computer could follow. The computer could be that expert. The machine would be able to diagnose disease, translate languages, even figure out what we wanted but didn’t know we wanted. 

AI got lots of  VC dollars thrown at it. But it was not much of a success.

Part of the (partial) failure can be attributed to a lack of computer processing power at the right price to accomplish those ambitious goals. The increase in power, drop in prices and the emergence of the cloud may have made the time for AI closer.

Still, I am not excited when I hear that this next phase will allow "services and advertising to be automatically tailored to consumer demands."

Gartner released a newer report on cognizant computing that continues that idea of it being "the strongest forces in consumer-focused IT" in the next few years.

Mobile devices, mobile apps, wearables, networking, services and the cloud is going to change educational use too, though I don't think anyone has any clear predictions. 

Does more data make things smarter? Sometimes.

Will the Internet of Things and big data converge with analytics and make things smarter? Yes.

Is smarter better? When I started in education 40 years ago, I would have quickly answered "yes," but my answer is less certain these days.


How Netflix Is Using Your Taste in Movies

With all the attention that privacy (or the lack of it) received in 2013, there are some forms of snooping that you might actually appreciate.

If you use sites like Gmail or Facebook, you probably know that they are mining your data and usage in order to give you ads that are better-suited to your interests. I know that may not sound so great, but it is an improvement on getting ads that are completely irrelevant to your life. But sites like Amazon and Netflix are also mining your data, not to show you ads but to show you more relevant recommendations. Their systems have become more sophisticated and more granular at judging your preferences. On Netflix, they look at the genres that you watch. In the early days, this would have been broad categories like drama, comedy, action, romantic etc. But their genres have gotten very specific, sometimes to a humorous degree - as in Fight-the-System Documentaries, Period Pieces About Royalty Based on Real Life, Foreign Satanic Stories from the 1980s.

Alexis Madrigalis a senior editor at The Atlantic, where he oversees the Technology Channel. He's the author of Powering the Dream: The History and Promise of Green Technology. He wondered how Netflix with its 40 million users (more than HBO now) decides on the genres that a film fits into in order to quantify your personal tastes.

We sometimes call this taxonomy or folksonomy, when it is done by "the crowd."

His interest turned into a bit of an obsession and then he discovered that he could scrape (capture) each and every microgenre that Netflix's algorithm has ever created. He discovered that Netflix possesses not several hundred genres, or even several thousand, but 76,897 unique ways to describe types of movies.

You may not be a movie fan, Netflix subscriber or even very interested in Big Data - but organizations (companies anf colleges) are very interested in knowing about you. Therefore, you should have some interest and understanding of what is being done to you.

Madrigal wrote a script to pull that data and then spent several weeks understanding, analyzing, and reverse-engineering how Netflix's vocabulary and grammar work. He realized that there was no way he could go through all those genres by hand, so he used a piece of software called UBot Studio to incrementally go through each of the Netflix genres and copy them to a file. 

He discovered many very specific genres in the system, such as:

Emotional Independent Sports Movies

Spy Action & Adventure from the 1930s

Cult Evil Kid Horror Movies

Sentimental set in Europe Dramas from the 1970s

Romantic Chinese Crime Movies

Mind-bending Cult Horror Movies from the 1980s

Time Travel Movies starring William Hartnell

Visually-striking Goofy Action & Adventure

British set in Europe Sci-Fi & Fantasy from the 1960s

Critically-acclaimed Emotional Underdog Movies

Perry MasonIn the article he wrote forThe Atlantic, there is a generator which will give you many of the genres. It is an imperfect system. He found an oddly large number of genres for the actor Raymond Burr (best known for an old TV show Perry Mason). Why? 

He explains: "The vexing, remarkable conclusion is that when companies combine human intelligence and machine intelligence, some things happen that we cannot understand. Let me get philosophical for a minute. In a human world, life is made interesting by serendipity," Yellin told me. "The more complexity you add to a machine world, you're adding serendipity that you couldn't imagine. Perry Mason is going to happen. These ghosts in the machine are always going to be a by-product of the complexity. And sometimes we call it a bug and sometimes we call it a feature. Perry Mason episodes were famous for the reveal, the pivotal moment in a trial when Mason would reveal the crucial piece of evidence that makes it all makes sense and wins the day. Now, reality gets coded into data for the machines, and then decoded back into descriptions for humans. Along the way, humans ability to understand what's happening gets thinned out. When we go looking for answers and causes, we rarely find that aha! evidence or have the Perry Mason moment. Because it all doesn't actually make sense. Netflix may have solved the mystery of what to watch next, but that generated its own smaller mysteries. Sometimes we call that a bug, and sometimes we call it a feature."

Gartner's Trends List for 2012

Now that we are midway through 2012, the Gartner Symposium gives us the IT research firm's tech trends for the year. As usual, it's a mix of emerging and existing technologies.

According to, the top ten includes:

    The use of media tablets and other small-form-factor computing devices;

    The continuing explosion of mobile-centric applications and interfaces;

    The growth of app stores and marketplaces;

    Contextual and social user experience;

    The "internet of things";

    Next-generation analytics;

    The proliferation of big data;

    The smarter use of in-memory computing;

    The recognition of the value of extreme low-energy servers; and

    The continued acceptance of cloud computing.

Technologies on the Horizon That Will Impact Higher Education

Every year I read the the NMC Horizon Report to see what they predict will be the technologies that will have an impact in higher education. The 2012 report was released jointly by NMC and the EDUCAUSE Learning Initiative.

The report looks at technologies that will have an impact in the next five years (near term, mid-term, and longer term). They also examine "critical challenges" facing education.

The near-term technologies are mobile apps and tablet computing which are changing the nature of computing for end users and developers. The larger suites of integrated software are being replaced by free and cheap apps that focus on doing one or a few things well and integrate with other apps easily. And, though I agree that mobile computing, tablets (iPads etc.) are influencing teaching and learning, I don't see any clear impact yet.

Two technologies that are more mid-term (2 or 3 years from having a major impact) are game-based learning and learning analytics.

"Learning analytics" may have more of an impact on the administration and decision-making levels than directly in the classroom. The term usually refers to both traditional strategies used in student retention, and newer methods of aggregating data from many sources to get a picture of how learning is happening and what is working best. If you have read previous reports, you know that long-term items, like learning analytics, often move closer in time if they grab traction in schools. Learning analytics, for example, seems to have benefited from some funded initiatives in the past few years.

Game-based learning has been on the list for a few years, but I don't feel like it has gotten any closer to making an impact. At one time, virtual worlds was a somewhat related technology, and that has almost dropped off the educational planet the past two years. Both technologies are ones that offer the possibility of using collaboration, problem solving, communication, critical thinking, and digital literacy. But the results have not been all that impressive. Online social games have certainly been big the past few years, but their application or any transference to learning is still lacking.

If you're writing your proposals for grants and conferences, you might want to get a jump on those technologies that are still four or five years out. Two to look into are gesture-based computing and the "Internet of Things."

Gesture-based computing fits right into gaming and mobile devices. Think of Wii games and swiping that smartphone or tablet.  The ideas driving its use in education is that it can transcend linguistic and cultural limitations. Watch a two year-old play with an iPad and you realize that not relying on language or any specific language might be a major plus. These devices also encourage interaction and just plain old play as a way to explore and learn. That is certainly true with younger students, but not lost on older and adult learners. Android and Apple smart phones and tablets, the Microsoft Surface, ActivPanel, Nintendo Wii and Microsoft Kinect systems, are all playing with these ideas.

Internet of Things

The "Internet of Things" is further out there in years and in my ability to explain exactly what it means, or might one day mean, to education. It is about the evolution of smart objects which are interconnected items in ways that make the line between the physical object and digital information very blurry or invisible.

You should look into IPv6 and how it is used in small devices with unique identifiers. You probably know a bit about RFID devices that are used in stores to track products, purchases and inventory. They store data and they can send that information to external devices via the Internet. We can already use them in schools to do similar things like tracking attendance, research subjects, and equipment. But how it might be used for learning is about as blurry as the line it is erasing.

Which brings us to challenges. In brief, these are the five technology-oriented challenges facing higher education according to the report.

1) Economic pressures from new education models, forcing traditional institutions to control costs while maintaining services;

2) The need for new forms of scholarly corroboration as traditional peer review and approval become more and more difficult to apply in light of new methods of dissemination;

3) The growing importance of digital literacy and lack of digital literacy preparation among faculty;

4) Traditional institutional barriers to the adoption of new technologies; and

5) Technological upheavals that are putting libraries "under tremendous pressure to evolve new ways of supporting and curating scholarship."

In my educational world, economics is very important, but the barriers of 3 and 4 are much tougher to overcome.