Conversation

Do you know if there's some work on this in the literature you've been exploring? It's naturally at the root of any training program design, so I suspect there must be?

Quote Tweet

Cortex Futura

@cortexfutura

Aug 3, 2021

Aside from KhanAcademy, what are some good illustrations of knowledge dependency graphs? I.e. "before you can understand Y, you need to understand X"

Replying to

Actually what I'm finding in the NDM/cognitive systems engineering literature is the opposite: they don't pay attention to knowledge dependency graphs but instead do tacit knowledge extraction from experts directly. Then they build training programs on top of those mental models.

Replying to

and

This is one of the 'bombshell implications' of the screenshots I posted yesterday — knowledge dependency graphs often do not reflect what is in the heads of actual experts, so incremental complexification of training tends to not accelerate expertise.

Replying to

Hm, very interesting! Might this be a question of level? I.e. for a novice programmer you have to introduce variables before talking about refactoring, but for experts that's taken as a given?

Cedric Chin

@ejames_c

Replying to

@cortexfutura

Yeah at first I thought that they were aiming for “how do we get journeyman level to expert level” type training programs. (Expertise acceleration). But they also assert that these methods may help accelerate novices getting to a min proficiency level. So I’m not so sure.

6:53 AM · Aug 3, 2021Tweetbot for iΟS

Replying to

Very curious, will have to read more! From my own experience I can say that there seems to be some sort of "complexity threshold" below which I *really* benefit from very clear, stepwise progression.

Replying to

I think the danger they warn against is that often, the mental models built from simplified lessons are often slightly flawed, and must be broken to reach higher levels of expertise. Given their goal is to accelerate expertise, they argue: skip this.

Show replies