Machine Learning: Reflections on the ATD TechKnowledge Conference

I had the honor to speak at the ATD TechKnowledge conference last week in Las Vegas, NV. This is a reflection on things you might find useful or at least interesting as well. Warning: you may need a good cup of coffee to get through this, as the post is longer than usual.

Lesson #0: It’s all about perception…

I did some research on gamification. The design of these machines is amazing. The sounds and visuals perfectly make you think you’re winning. A penny is not 0.01 but 1. So, you’re so excited when you win 50!! Yet, it’s 50 cents. And your bet was actually 150 cents. You see the trend here… IMG_34523 


Lesson #1: (Deep) Machine Learning is not about Terminator 5

David Rose was talking about Enchanted Objects and Machine Learning in his opening keynote speech. While this picture refers to products in general, don’t forget that learning solutions are products over all! And those steps are SO relevant to L&D!! Let’s move from order taking to great experiences in 2016!


Machine Learning? Over 20 years ago, I was working on my thesis. I built an artificial neural network in C++ that was supposed to learn how to add numbers together. Basically, imagine a black box. You show two numbers to the box. Then a box works its magic inside and spits out the sum of the two numbers. It’s WAAY off… When you show the actual result, the black box tweaks itself and you go again. Show another two numbers. Black box spits out its guess. Still off, somewhat closer. Adjust. Tweak. Show again. And this is going on and on with tens of thousands of examples… Until, at some point, the system learns what it means to add two numbers together. It’s still not 100% all the time but pretty good. My network was running for weeks to learn a simple thing a first grader can do. I graduated. That’s what counts.


Today, we call this machine learning. Adding numbers together might not sound cool (although, I tried to apply my thesis to lottery tickets to see if it can spit out the winning numbers) but imagine the same for face recognition. David Rose showed us the results how marketing is using machine learning. Do you want to know how and when your product is used? It’s not about focus groups anymore. It’s the real thing. Machine learning is about telling the system what you’re interested in and let the system figure out how to find the information. Imagine manually sifting through billions of pictures posted daily on social media and find who’s eating ice cream, where and in what context. What’s the trending flavor? What are they wearing? Are they happy? We’re focusing on “The Big Picture.” Literally. Our brain does this all the time: we recognize a number upside down, twisted and half torn. How? Who knows? Our brain can naturally recognize the pattern that makes a number a number.  And that’s the key here. Machine learning is not about machines taking over learning. It’s about machines applying “simplified” pattern recognition algorithms imitating the brain. We, humans are too close to the big picture and do not have the capacity to see the patterns, only a pixel or two. Machine learning is a trend to keep an eye on!


Lesson #2: 1 in 6 people hate games the research says


My hands on session was called: Release Your Inner Angry Bird: Engaging Minigames in E-Learning. So, imagine my shock when I learned about this research that 1 in 6 people hate games. Who would come to my session then?

The research I conducted was based on my shirt. Randomly sampled a couple of times a day, then aggregated and exaggerated into an infographic. I can also call it a wearable object to be cool.

With all seriousness!! People! Just because “data make sense” on an infographic, it does not mean it’s based on research, OR that the simplified interpretation of the result is valid in our of context (reminds me of an article posted on LinkedIn where the person claimed that interactivity hinders learning. Except, when you dig up the actual study, you learn that “interactivity” was defined as clicking the next button to go through over 100 images, as opposed to see the same information as a continuous video.)

It’s always good to have a rounded picture of anything related to learning, especially if it’s something relatively new like gamification of learning (,, ) or neuroscience of learning (, The more articles you read, the easier you recognize patterns (machine learning) and the more confident (within a margin of error) you will be in your standing.

Because it’s not about where the truth lies! It’s about how confident you are (what is your acceptable margin of error) about where you’re standing on the line knowing what we know today (not just by reading your shirt).

Lesson #3: Curation vs. Creation

There’s actually a debate on whether instructional design, as we know it, will be around for any longer. If one thing was clear from talking many at the conference, was that curation of content and the shift towards real-time performance support are pressing issues no matter what field you’re in.
IMG_3449 Many of us facing a world where change happen so fast that traditional instructional design may no longer be agile enough to even catch up. If you’re not familiar with agile/lean startup methodologies combined with game thinking (, I would strongly suggest learning more about them. Traditional design is centered around content, while the world is now focused on the experience with the user in the middle (UX). People want products that work, not products that need some training. Does that mean we don’t create training anymore? No. It means we must not recreate the wheel (taking one form of content and put it into another), but curate already existing content and build action-based experiences around them. We simply will not have the time to create training when the problem is not even a training problem.


 Lesson #4: Gamification vs. Game-based Learning

If you know me, you know you should never ask me about gamification/game-based learning because you’re going to get stuck in a long and passionate conversation with me (like when Julie Dirksen accidentally sat down next to me at the Speakers’ Room). And this brings me to my session. The hands on session was about using a game engine (Construct 2). I designed my session to attempt a balance between hands on practical experience with an application (most people had no previous experience) and providing a vision and inspiration strong enough for participants to explore after the conference.


“game mechanics”

Previously, I attended Jane Bozarth‘s session on how to show your work. As a contribution, here’s some behind the scenes thoughts on my session design and delivery.

These were my design guidelines:

  1. I don’t want to do an application training only (click here, do this, do that).
    Why? Because in an hour you may learn where to click but you will never remember why. I can push you through some beginning steps but you’re going to need the pull when you return to work. You’ll need an intrinsic drive, a destination of your own, rather than my road signs and rules of traffic at the start line.
  2. I want to be clear that engagement/motivation is not about game mechanics thrown into a boring learning.
    Why? A game engine has all game mechanics built in at your fingertips. So now you can build an engaging game? No. It’s like saying that just because your computer has all the letters on the keyboard, now you can write an top selling novel.This is the picture I want people to take away from my session. I want them to use “Game Thinking” to resolve problems. Game Thinking may result in gamification, game-based learning, gameful design or none of these. Sometimes, it’s not even a training issue. The point is, do not start out with “I have a game engine, I’m going to build a game or gamify content” attitude. Forget game mechanics (points, badges, quests, etc.). Think about what the user should FEEL. The EXPERIENCE. Focus on the MVP (minimal viable product) that brings the experience to life and supports your ultimate PERFORMANCE objectives.Did you ever clients who signed off on your storyboard and when they saw the elearning for the first time they said that’s not what they were expecting? Even if you clearly indicated what will happen at each click? That’s the power of prototypes. Early and often. Do not rely on the brain’s “fill in the gap” feature that auto magically completes pictures.
  3. I want to provide resources that they can use immediately when returning to work. I guarantee people will forget where to click to open a new project. But they should not forget the one picture above with the “kid flying.”
    How? I built an interactive Inkling book with hours of content recordings, theory and people to keep an eye on. Resources include all source files and an integration package with Storyline. Yes, it basically took forever but that’s the only way some of the people who showed up for my sessions (over 100 ppl) will eventually use this.


“game dynamics”

  • How did the sessions go?

    Honestly, the first session was challenging. I had 59 participants for an hour of hands on learning, where many of the participants have never used a game engine, some just started to explore elearning itself. Showing the value and the potential of Game Thinking AND provide hands on experience with a game engine was not an easy task.

    Based on the first day experience (and the fact that I woke up at 3am the next day), I tweaked the session and adjusted the flow for the next one. It worked really well.

    Thanks for all who attended either sessions! I loved your enthusiasm and I do hope that my session gave you some inspiration on how to make learning more engaging and motivating.

LESSON #5: The Power of Social Media

Finally, one of the most important learning I take away from the conference. Technology and trends come and go. People are here to stay. We are not individual heroes inventing the wheel out there! The power of networking, online and offline is not an advantage anymore. It’s essential. Social media allows you to connect with the right people (who turn out to be actually nice in person as well) and exchange thoughts in real-time. You don’t need to wait until a conference to discuss any of these points above. You can do it today on LinkedIn or Twitter. So, I just want to thank everyone out there in cyberspace for what they’re doing whether it’s game design, gamification, learning or performance support. No. You don’t need to be connected 24/7. You do need some reflection time on your own but you can only reflect on things you learn (other than from your shirt). From others. Readers like you! Thank you!


“Reflection time in Red Rock Canyon, NV”



1 thought on “Machine Learning: Reflections on the ATD TechKnowledge Conference

Leave a Reply

Your email address will not be published. Required fields are marked *