Schemata - Blueprints Of The Mind

Quote:

Originally Posted by The Cusp

Schemas and AI

I was going to expand on Archetypes, Schemas, and Artificial Intelligenge to try to give a better understanding of the types of associations that are made when new detail is formed as a result of your attention. But it turn out Archetypes are a type of Schema, along with stereotypes, social roles, and worldviews. Of particular interest inception of the Schema concept.

The following article is from the Wikipedia entry on Schemas.

The original concept of schemata is linked with that of reconstructive memory as proposed and demonstrated in a series of experiments by Bartlett (1932). By presenting participants with information that was unfamiliar to their cultural backgrounds and expectations and then monitoring how they recalled these different items of information (stories, etc.), Bartlett was able to establish that individuals' existing schemata and stereotypes influence not only how they interpret 'schema-foreign' new information but also how they recall the information over time. One of his most famous investigations involved asking participants to read a Native American folk tale, "The War of the Ghosts," and recall it several times up to a year later. All the participants transformed the details of the story in such a way that it reflected their cultural norms and expectations, i.e. in line with their schemata. The factors that influenced their recall were:

* Omission of information that was considered irrelevant to a participant;
* Transformation of some of the detail, or of the order in which events etc were recalled; a shift of focus and emphasis in terms of what was considered the most important aspects of the tale;
* Rationalisation: details and aspects of the tale that would not make sense would be 'padded out' and explained in an attempt to render them comprehensible to the individual in question;
* Cultural shifts: The content and the style of the story were altered in order to appear more coherent and appropriate in terms of the cultural background of the participant.

Bartlett's work was crucially important in demonstrating that long-term memories are neither fixed nor immutable but are constantly being adjusted as our schemata evolve with experience. In a sense it supports the existentialist view that we construct our past and present in a constant process of narrative/discursive adjustment, and that much of what we 'remember' is actually confabulated (adjusted and rationalised) narrative that allows us to think of our past as a continuous and coherent string of events, even though it is probable that large sections of our memory (both episodic and semantic) are irretrievable to our conscious memory at any given time.

Further work on the concept of schemas was conducted by Brewer and Treyens (1981) who demonstrated that the schema-driven expectation of the presence of an object was sometimes sufficient to trigger its erroneous recollection. An experiment was conducted where participants were requested to wait in a room identified as an academic's study and were later asked about the room's contents. A number of the participants recalled having seen books in the study whereas none were present. Brewer and Treyens concluded that the participants' expectations that books are present in academics' studies were enough to prevent their accurate recollection of the scenes.

For starters, lets look at the factors that influenced recall.

* Omission of information.
First rule, without your attention, things can't exist

* Transformation of some of the detail; a shift of focus and emphasis in terms of what was considered the most important aspects of the tale;
Again, a combination of the first and second rule, and your most powerful tool in dream control.

* Rationalisation: details and aspects of the tale that would not make sense would be 'padded out' and explained in an attempt to render them comprehensible to the individual in question;
Second rule, attention creates detail.

* Cultural shifts: The content and the style of the story were altered in order to appear more coherent and appropriate in terms of the cultural background of the participant.
This sounds like the effect emotions have on dreams. Is it possible different cultures have an identifying emotional state?

These factors responsible for flaws inconsistencies in long term memory just happen to be the tools of dream control! The links to memory make me wonder if dreams are really dynamic memories kept alive by our attention.

And now for an Artificial Intelligence approach. A binary yes or no system may be easier to picture than fancy notions like Schemata.

Consider training an artificial neural network to understand language, starting with a concept like tree. You could manually create links to words like grows, leaves, branches. Alternatively, you expose the network repeated instances of trees until it learns on it's own. Links to words like outside would be formed, but not all instances of trees would necessarily be outside, some may be indoors, so another link to indoors is formed. But the majority of trees being outside, the path to "outside" gets used more and gains weight, becomes more substantial. The more connections or paths your word has, then better the understanding your AI will have.

You have to learn to travel those weighted paths or synapses without getting lost within their endless nature.

lol