Top Page | Upper Page | Contents | About This Site | JAPANESE

Cognition and Learning

Cognitive psychology and brain science study how humans perceive.


It means "typical example", "type" or "pattern". Because of this, even if detailed information is not enough, it seems that you can understand what the other person is saying.

Once you learn the schema, you will be able to apply it as an analogy or an abduction .

Rustic concept

It is a model of natural science that we know empirically. For example, when playing baseball, you can predict where the ball will fall, move around, pull out your glove and catch it without knowing physics.


The range when thinking about something is called "frame", and defining that range is called "framing".

For example, when we are talking about "I was the best", we unconsciously remember when and where, and this is the "frame". What happens to the frame is said to be "framed" depending on the situation before the conversation begins.

Framing effect

Framing is often done unconsciously, so it is very difficult to notice. But the effect can also be a reason for a person's actions.

The framing effect is quite conscious in questionnaires, sensitivity evaluation , and behavioral economics .

Learning patterns

"Humans do not think about things using all the information down to the smallest detail, but judge patterns from limited information and take action according to the patterns ."It can be read from psychology-related fields such as science, psychology of decision-making, and heuristics .

The word "pattern" means the same thing as the words " schema ", "type", "style", "typical example", "empirical knowledge", "rule", "formula" and "feature". there is.

For human learning, it is important to memorize patterns in the world through repetition of real-world situations. By remembering patterns, you will be able to recognize things smoothly. By the way, a neural network is a mathematical application of the way the human brain learns.

The part where the pattern learning is not done well becomes "fixed idea" or "belief". Thank goodness for stubborn people. In order to get rid of "fixed ideas" and "beliefs", it is necessary to review and update the accumulated patterns, but I think it takes courage to update the patterns.

Range of human perception

It is very difficult for humans to recognize things other than patterns. Humans usually do not perceive the whole of things. Especially when you need to make a quick decision, it's even more difficult.

This is one of the foundations of the human imperfection argument in behavioral economics and the imperfection argument of people who use science.


Thinking about "meaning" is a very important approach, and it is necessary in both product design and urban planning . In the technical field, we are actually doing things like reading meaning from patterns and putting meaning into patterns when designing.

"Pattern", "Recognition" and "Meaning" are closely related. Recognize patterns to know their meaning, or express them in patterns to make them recognize their meaning.

Language learning

In philosophy and other fields, there is a hypothesis that "human language recognition is determined by learning examples." In fact, current recommendation systems and natural language processing are based on this concept.

However, when considering the method of language acquisition in Applied Behavior Analysis (ABA) , it is not just "learning from examples", but also learning through things that actually exist. It seems to be good to distinguish this distinction and usage according to whether the word is abstract or concrete. The figure below summarizes this idea.

Application to artificial intelligence (AI)

Machine learning learning is not the same as humans, but there are similarities.

Utilization of Perceptual Research

Looking at an image of an apple and judging it as an "apple" is a method called supervised learning in machine learning.

Language learning

In machine learning called "natural language processing," some use dictionaries, while others allow computers to learn the structure of language and the meaning of words on their own.

As for how it learns, it learns from a huge amount of usage examples, such as the words before and after the word, and in what field it appears.

This also enables automatic statements .

The point is that even if the meaning of a word is not fixed somewhere, the meaning can be acquired from other words.

Learning and Leveraging Schemas

The method of learning a schema is reinforcement learning in machine learning.

Association analysis is also a way to learn many schemas at once. Used as a recommendation system .

NEXT Imitation and Counter-imitation