 |
|
|
 |
|
|
|
|
Hamun |
| | | | |
tkorrovi |
| | | | | | posted 2/11/2012 19:03 |   | | | | | | | | | | |
Hamun wrote @ 2/11/2012 6:18:00 PM:
learning is gathering of information and more developed using this information,so the robot can make a self programming
| |
Oh, youtube videos and colorful presentations, they take my breathe away. Like a shampoo advertisement, after seeing it i feel that this particular shampoo must be good for me. It was very difficult for me to watch this long video with my slow internet connection.
No, learning is not only gathering information, it is building models based only on the information received from the senses and interacting with the environment. A process which needs a lot of flexibility. This is awareness and yes awareness is action, because it is about processes, it is perceiving processes. This is what i have told here already for many years.
And it is not self-programming. Generating code is not a solution for things like these, because code in whatever programming languages contains too much excess information, so the generated code would most likely not work. Self-development is what we need instead. And self-programming or generating code was also not what this video talked about, and it was not said to be a way of doing things there. This video said that the perception process itself is the action, the same which i have said here, which confirms that this is the way it is.
| | |  Artificial Consciousness ADS-AC project | | | Last edited by tkorrovi @ 2/11/2012 7:05:00 PM | | |
| | |
Hamun |
| | | | | | posted 2/11/2012 19:54 | | | | | | | | | | Thank you for your response.
Actually is not learning is from our teachers and from trial and error.
| | |
| | |
tkorrovi |
| | | | | | posted 2/11/2012 20:36 | | | | | | | | | | | |
Hamun wrote @ 2/11/2012 7:54:00 PM:
Actually is not learning is from our teachers and from trial and error.
| |
No, an aware entity supposed to learn also only from the objects in the environment. I have used this example there. Put a truly aware robot on wheels in a room where are some cubes on the floor. It starts to move by itself, trying to find increasingly more difficult routes over the cubes, thus exploring its environment and itself.
And this is not just trial and error, because everything new is developed based on what has already been learned, deriving the new from the old by self-development. There are of course still enormously lot of trials, but within the constraints of what is already learned. Also because of these trials such systems need a lot of processing power, a lot more than pre-programmed systems which take action by fixed rules, so pre-programmed systems may seem to be much more functional, which in fact is not true because of their enormous disadvantages in dealing with novel or unexpected situations.
| | | Artificial Consciousness ADS-AC project | | | Last edited by tkorrovi @ 2/11/2012 8:48:00 PM | | |
|
|
|
|
|
