Principle that the learning curve is an S-curve
https://gyazo.com/cc5eda93b5e5b7ccc1bdac3439dc9f03
Model the knowledge acquisition process as follows
many-to-many
multistory
Upper layer knowledge cannot be acquired directly.
When you acquire lower layers of knowledge, you acquire higher layers of knowledge by luck #Abstraction The more you know about higher layers of knowledge, the lower the cost of acquiring lower layers of knowledge.
In this model, the amount of knowledge gained per unit of time is [graduated
Because the cost of acquiring knowledge at the higher layers gradually decreases and the cost of acquiring knowledge at the lower layers decreases, increasing the amount of knowledge acquired per cost.
On the other hand, the number of upper layer knowledge per number of lower layer knowledge gains [saturation
When 0 → 1, any of the N higher layers of knowledge may be acquired, but when N-1 → N, the last one must be acquired, since there is an approximately N-fold difference in probability
These combine to form an S-shape curve. ---
This page is auto-translated from /nishio/学習曲線がS字曲線になる原理. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I'm very happy to spread my thought to non-Japanese readers.