Is adaptive learning something you can leverage science and technology (S&T)? Do you work with Office of Naval Research (ONR)?

We work closely with the ONR. We do lots of work with ONR. One of the ways is adaptive learning. For example, the intelligent tutoring is a category under adaptive learning and it’s more aligned to the classroom, the book-type learning. The intelligent tutor system learns about the student and then adapts the training syllabus for that person.

So there’s a program with ONR right now that our NAWCTSD folks are working on called ATSNAP – adaptive training submarine navigation and piloting. It’s very much what you just described as an example. The submarine conning officer can look into a periscope of a submarine and try to identify ships and determine their range and angle on the bow. So when it recognizes that the person should have seen the range and the angle on the bow, much earlier, and it gives some amount of feedback. But if the person got it right, its take them to the next level, then the next and so on. They call it “leveling up” in gaming. The person who doesn’t learn quite as quick gets some level of additional instruction; and the person doing well is successively challenged more quickly. This is where the science of learning plays in. These are just small examples of how to use adaptive learning. The intelligent tutoring is specifically talking about a classroom environment, less in a simulation but more in a classroom.

We work closely with the ONR. We do lots of work with ONR. One of the ways is adaptive learning. For example, the intelligent tutoring is a category under adaptive learning and it’s more aligned to the classroom, the book-type learning. The intelligent tutor system learns about the student and then adapts the training syllabus for that person.

 

How do we assess someone in a meaningful way? How do we build a training system that can do that and then what do we do with the kind of metric? And how do we leverage all that to determine trends, or use that as a management tool?

NAWCTSD uses the assessments to improve our training capability to make the training systems better.  The primary goal is to get that student and that crew in train to train to qualify/train to certify, to the right level, based on what the fleet requirement is. And we use that information, on how that person is learning, to adapt the training to get that training better and better and better.

That’s what we call training optimization. Our research psychologists are continuously looking at that. So is  Vice Adm. Dunaway, and not just for aviation products, but here at NAWCTSD we do all these other areas of the Navy, too. Training optimization means that we look at the data that we have.  For example, in the aviation pipelines we send people into a flight school. And they learn generic airplane flying. They learn to fly the T-6 Texan, and the T-45, and they roll up through there and they get their wings, and it’s almost like a train to qualify up to that point. And then they flop over and they get their aircraft assignment and go on to fly the F/A-18, and they fight the F/A-18 as an individual, then maybe section, then division, then air wing, and then coalition and, you know, that whole thing. It may be possible, potentially, to do some of that even in the flight school part, where we do some of that T-6 Texan flying, and T-45 flying, some of those syllabus events, which can be in the less expensive airplane, maybe, than the more expensive airplane. And move that syllabus around by looking at the science of learning and how people have performed along the way.

You look at those exact objectives and how well they’ve learned those and how expensive or slow the training system, or the training media selection was, “media” meaning like a particular airplane model or in a classroom, or immersive simulation. Can we move events from the airplane into the simulator? Can we move it from one airplane to a different airplane? Can we move it to a classroom? Would that make the overall training syllabus shorter? Would that improve cost and better students? Would it make it longer?  That’s training optimization.  That’s what we would use for LCS and train-to-qualify and train-to-certify, also.

 

And the requirement that you’re training to comes from the fleet.

Yes.  They say, “this is exactly what I need on a LCS crew member.” The aviation community has been doing this for a little longer, so they’re much more detailed in their requirements. They have what we call a “training and readiness matrix,” T&R matrix, and it says exactly what the person needs to be able to do, for example, in air to air combat; or air to ground missile employment; landing on an aircraft carrier  — what all those requirements are, and then what they need to do, and how they need to do it.

The aviation community has been doing this for a little longer, so they’re much more detailed in their requirements. They have what we call a ‘training and readiness matrix,’ T&R matrix, and it says exactly what the person needs to be able to do, for example, in air to air combat; or air to ground missile employment; landing on an aircraft carrier  — what all those requirements are, and then what they need to do, and how they need to do it.

 

Such as “Launch an F/A-18; conduct a mission; land on a pitching deck at night in bad weather?” How do you do that?

If we were just buying things on contract, we’d say, “okay, just train me to do that.” But what we’re trying to do with our science and learning people is determine how is the best way that a brain really grabs onto that and knows intuitively when it really is a worse situation: pitching deck, bad weather, night time, single engine on the F/A-18; how is he going to be able to do that? And will he do that, in fact, in the fog of war, fight his airplane onto the deck. What are all the requirements of that brain to really get that?

How do we do media selection based on that? If it’s some adaptive training situation, how do we adapt that training as the person is showing that they’re getting under power and close to the back end of the ship, at the wrong time, when they’re single-engine? We look at it altogether and say, “This is the smartest way to do it.” Sometimes that’s too expensive for the resource providers to buy. But at least we have the data to show how the brain works, and the science and learning behind the brain that says, “this was the right way to do this.” If we can’t do it for some reason, okay, then we’ll some alternative, another way to do it. But, realize that there will be a little higher risk on that person on that night, and you have to make sure that you have a bunch of smart people on the radio talking to him, and the landing signals officer on the flight deck knowing if he gets a little under power, he might crash that airplane, so get ready to call “eject” or something.

There are so many things that can go wrong in an airplane that your brain is continuously reviewing everything, prioritizing and deciding what to do next.  In aviation, they say the three top priorities are: aviate, navigate, and communicate – especially when there are any casualties or emergencies going on.