Learning trends

There’s nothing like predicting your learners’ needs

Read in 4 minutes Modified on
Predictive learning

Predictive learning is a system of machine-based learning that estimates changes in the environment, makes these changes in the learning program, and allows employees to learn how to deal with the different surroundings.

Predictive learning’s opposite is Representation learning, showing things the way they actually area. The concept works on the very logical assumption that learners may be learning new skills today, but they will be using them tomorrow when the action environment will have changed, at least to some degrees.

It fits in with the idea of flexibility or agile learning, discussed elsewhere in this blog. A good predictive learning program takes into account that no matter how well designed the program, learners are likely to encounter situations the program did not anticipate.

What place for predictive learning?

Predictive learning fits nicely with a general learning program that uses different methods of training, necessary for individualized learning.

Predictive learning takes advantage of the nature of learning on an electronic device. It uses the ability of electronic on-line learning programs to frequently change. Stated conditions can change at the start of a program. They can change during the program to reflect leaner actions :

This approach can make unannounced changes, to be used to accustom learners to unpredictability. It helps develop learner flexibility, a “thinking” skill every bit as important as functional skills.

This is a concept that, in its basics, long predates mobile learning. Military training, probably back before the Romans, tried to predict how an enemy would act and find ways to deal with the actions. The modern military has created the “red team,” to play the role of the enemy in war games. The red team is under no obligation to avoid changes in strategy and methods.

Learners can develop different ideas on how to react to the changing situation they are given in the programs. This training can even take into account unexpected changes. Trainers can guess on what to interject into the training. They can make training totally unpredicted or what is called a “black swan”, a seeming surprise which should be predictable.

Predictive training can even include looking at what might change and how learners might respond.

An advanced computer and a 75-year-old manual typewriter have certain things in common. The basic result is the same, words on paper as another transmission medium.
The actual typing is the same on both. But most people know to take advantage of the many differences in the two systems :

So how might a program be designed to let learners deal with different price levels for oil, to give a currently relevant example?

Limit time to make decisions

The first element might be to allow limited time to make decisions. The system can put absolute limits on time. It might also give different result on decisions made “now” or at a future time.

A good learning designer can start by asking the students to determine how long they have to make a decision. Real life decisions might have an imposed time limit, but not even emergency decisions, outside of a TV suspense program time bomb with a countdown clock, tell you exactly when a decision has to be made.

A realistic training program can ask learners how much time they have to make a decision, but not honor student prediction. Decision makers often do not have as much time as they think they have to make decisions.

The timing decision is the first question of the program. Tell the students they have to determine how long they have to make an initial decision.

Decision timing is the first decision to make in virtually every decision making. And, like very decision, it has consequences. The later phases of the program can filter these consequences.

Offer decision-making at different times during the training

As in real life, decision points in a training program can occur at different points of the program. Instructors can recommend the best action in certain situations, but must also be sure learners understand that there often a continuum of “accurate” answers.

Actions have different results; and answers are often not right or wrong. Actions have consequences, which differ according to the nature and timing of the action. Predictive learning enables the program to change as the learner responds.

These programs can also, on occasion, include situations where the what seems like to the correct decision leads to the wrong result. A really good program can ask the students – what happens? Why did it happen? Could you prevent it ? How do we move along from here?

Adapt to the learner to predictive learning

One more thing. A well designed predictive learning program can train learners to deal with different and changing situations, with changes depending on learner actions – as in real life. But, like any training program, learners will have to decide when particular skills and actions are appropriate.

Not all actions work every time. Not all training works for every situation. Perhaps the key skill to develop in a training program is flexibility.