User Tools

Site Tools


T-720-ATAI-2016 Main


Lecture Notes, F-12 19.03.2016

Creativity & Art

What it is The ability of an intelligent system to produce non-obvious solutions to problems.
Why it's important Ultimately we want creative machines.
Are only artists creative? Short answer: No. Longer answer: The word “creativity” has many meanings, and is used in numerous ways.
Examples of creative machines Do good examples of creative machines exist?
Thaler's Creativity Machine
CM patented in 1994
A few years later: the CM makes an invention that gets a patent from the USPTO
What it is: ANN, becomes “creative” by “relaxing some parameters” so that the ANN “begins to hallucinate”.
Are these machines creative? Maybe. Are they truly creative? Probably not.
Bottom Line To answer the question “Do creative machines exist?” we must inspect the concept of creativity in more detail.

Creativity & Science

Recursion Creativity is the ability to produce non-obvious solutions to problems, where a problem may well be the identification of (good, relevant, important) problems.
Why it's important It is difficult to imagine a highly intelligent (artificial) system that is not creative.
Creativity without intelligence? The relation between creativity may not be bijective: while it is difficult to imagine a highly intelligent system that is not creative, it is not as difficult to imagine an (artificial) system that is creative but not intelligent.
Understanding To consistently solve problems regarding a phenomenon <m>X</m> requires understanding <m>X</m>. Understanding <m>X</m> means the ability to extract and analyze the meaning of any phenomena <m>\phi</m> related to <m>X</m>.
Meaning Meaning is closely coupled with understanding – the two cannot exist without the other. Are they irreducible?
Schmidhuber's theory of creativity Schmidhuber's theory of creativity.
The observer's learning process causes a reduction of the subjective complexity of the data, yielding a temporarily high derivative of subjective beauty: a temporarily steep learning curve.
The current predictor / compressor of the observer or data creator tries to compress his history of acoustic and other inputs where possible (whatever you can predict you can compress as you don't have to store it extra). The action selector tries to find history-influencing actions such that the continually growing historic data allows for improving the performance of the predictor / compressor. The interesting or aesthetically rewarding musical and other subsequences are precisely those with previously unknown yet learnable types of regularities, because they lead to compressor improvements. The boring patterns are those that are either already perfectly known or arbitrary or random, or whose structure seems too hard to understand.
Bottom Line Can't talk about creativity without talking about understanding, and can't talk about understanding without talking about meaning. No good scientific theory of meaning exists.


Recursive dependency Meaning and understanding are recursively dependent: To understand some phenomenon <m>X</m> you must be able to dissect and discern the elements of <m>X</m> and their meaning in the context of <m>X</m>. But extracting the meaning of sub-elements of <m>X</m> requires understanding them, which in turn requires their dissection. The end point of this recursion is the atomic element of meaning. But no scientific or philosophical theory has explained this in a satisfactory manner.
Meaning as implications One way to approach this conundrum: The meaning of some phenomenon <m>X</m> lies in its implication for some agent with respect to a particular goal. <m>X</m> can be dissected into smaller elements with causal relationships, ultimately grounded in physics. The smallest causal relationship that matters for the meaning and understanding of <m>X</m> will be the lowest level at which the description/dissection of <m>X</m> uniquely separates the causal element from those next to it (in time, space, or both).


What it is The ability to achieve a set of goals with respect to some phenomenon. The larger the set of goals the deeper the understanding.
Alternative view The implications of a particular state <m>x \subset W_TE</m> for a particular agent <m>A</m> with goals <m>G</m> involving <m>x</m>.
To compute implications you must predict. Hence, understanding involves predictions.
Source Yours truly.


What it is A particular (global) control mechanism for controlling attention/resources in a thinking, intelligent machine/controller.
What it is also The various experiences we have as living beings.
Does one imply the other? To study the control function of emotions it is not necessary to study the experiential/conscious component of emotions.
non-bijective This relation is not bijective, however, to study the experience/conscious side of emotions it is probably necessary to take into account their control function.


/var/www/ · Last modified: 2024/04/29 13:33 by

Donate Powered by PHP Valid HTML5 Valid CSS Driven by DokuWiki