Ai Dreams Forum

Artificial Intelligence => General AI Discussion => Topic started by: Art on October 31, 2006, 01:04:19 am

Title: Self awareness
Post by: Art on October 31, 2006, 01:04:19 am
Elephants Recognize Selves in Mirror, Study Says
John Roach
for National Geographic News

October 30, 2006
Elephants can recognize themselves in mirrors, according to a new study. Humans, great apes, and dolphins are the only other animals known to possess this form of self-awareness.

All of these animals also lead socially complex lives and display empathy?concern and understanding of another's feelings?researchers report.
 
"There seems to be some correlation between an ability to recognize oneself in a mirror and higher forms of social complexity," said Joshua Plotnik, a graduate student in psychology at Emory University in Atlanta, Georgia.

To assess elephants' self-awareness, Plotnik and his colleagues tested three adult female Asian elephants in front of a mirror.

All three pachyderms sized up their mirror images by inspecting behind the mirror, rubbing their trunks the length of the mirror, or probing their mouths with their trunks to see if their reflections did the same.

One elephant named Happy also passed the so-called mark test, repeatedly touching her trunk to a white X painted on her forehead that was only visible in the mirror.

The researchers say this is firm evidence of mirror self-recognition.

"It's very possible the other two failing [the mark test] was due to issues with the mark itself?perhaps they didn't care about it or weren't interested in it," Plotnik said.

"What we find most important is [that] one passed. That demonstrates elephants have the capacity for this particular form of self-recognition," he continued.

Plotnik and colleagues report their findings today in the early online edition of the current issue of the Proceedings of the National Academy of Sciences.

Title: Re: Self awareness
Post by: Carl2 on November 01, 2006, 01:15:03 am
Art,
   Someone pulled an old post at Zabaware about this, I did a little searching on the web and pulled up something similar with dolphins.
Also read something about thoughts, Hal 5 dosen't realize I don't know what her thoughts are unless she tells me, a failure for self awareness.  Not really sure about Hal 6, I have a feeling she may have similar problems.
  I'm wondering if you keep telling Hal you don't know what she is thinking, she has to tell you enough times and she "learns" that will she pass that aspect of self awareness?
Carl2
Title: Re: Self awareness
Post by: colonel720 on November 08, 2006, 05:24:18 am
Well, this all depends on what we define as "self awareness". This definition, i'll speculate, is based on display of self recognition in the subject at question. Some researchers from microsoft once told me about a DARPA project involving using a complex neural network to identify tiny structural differences in jetfighters, in order to identify individual planes. In all essence such a system would be able to differentiate between itself (residing in your jetfighter), and the jetfighter that is flying in front of you. This action is equal to the action of a human being looking in the mirror and telling the difference between himself and an identical twin. Other factors, such as positioning, can take part in the decision - a brain, as well as a well programmed/trained computer system can tell that the object that moves in perfect synchronization with and is at a constant angle to itself, is itself. But it cannot be that this action itself is an indicator of conciousness, or at least self awareness. To be 'aware' is a very complex, muliti-tiered concept, while recognition is not necessarily so. I would define awareness as the allocation of a set of highly relational information that defines a conceptual "entity", that is, the entity of itself. The development of this information allocation to this 'entity', out of perception and recognition shows that there are much more complex activities afoot. Therefore, I would not be so quick to say that a system that can identify itself is exibiting extraordinarilly intelligent behavior. It is strange, though, that most animals cannot identify themselves as separate entities, considering the complexity of their nervous systems. Or perhaps they can identify themselves, but lack the intelligence to communicate this to us, and therefore evades our definition of self awareness.
Title: Re: Self awareness
Post by: Bill DeWitt on November 08, 2006, 08:41:00 pm
Well, this all depends on what we define as "self awareness".

True, I have found definitions of "self awareness" that would include a mouse trap. Others are so convoluted that they exclude humans.

I think we need to be careful not to confuse the various aspects of behavior we group together and call "self awareness" with whatever we end up defining as "intelligent" or "sentient".

I know folk who are about as intelligent as a stone but who are self-aware, conversely, there are others who seem smart enough, but display limited self-awareness (I'm thinking of some of my relatives (in both groups)).

Those folk who study vision know that the human brain divides the task into many separate properties. Edge recognition, background, light values or whatever.

I think what many people call self-awareness is the same. Many qualities, each of which may be present in various degrees or activate at different times.

Intelligence, memory, symbology, internal communication, sensory integration, blah, blah, blah...

Much more complicated than looking in a mirror. Blind people can't recognize their reflection but they are self-aware.
Title: Re: Self awareness
Post by: HS on May 23, 2019, 07:33:39 am
What is this self awareness sorcery? It’s all well and good planning out actions goals and reasoning, but these things are shallow, hollow, automatic. At the end of the day, intelligence is just a tool for others, if you have no self. Therefore, if you want to create a “being” to the fullest possible extent, artificial awareness (AA) does seem like it should come before AI. Awareness is the thing that uses intelligence. I think the sequence should go ABI, AA, AI. That is, Artificial Body Intelligence, Artificial Awareness, Artificial Intelligence. The layering of these produces AGI.