Slashdot Mirror

Stephen Thaler's creativity machine is proof of the potential of machine creativity.

The improvement has been using multiple ANNs together as communicating units that can both communicate information and dynamically train each other, instead of trying to make a single large ANN and using external training sets. Of course, this isn't that new, as these guys have been working on such models since at least the early 90s.

http://video.google.com/videoplay?docid=-6464697696665901632&ei=hEpzS-jkIpPorAL02JD-Aw&q=initsimage+google+video&hl=en&client=firefox-a# You just don't know about it. I'm sure this same technology has been pushed much further by the recent advances in processing power. http://www.imagination-engines.com/cm.htm

Hawkins' solutions likely overlap with Dr. Stephen Thaler's patents for neural networks(NN). In particular, Thaler's algorithms inject noise into a proprietary NN system (actually 2 or more NNs conjoined) to generate novel patterns (that is, to _discover_ new patterns). For example, Thaler trained and used such a NN system to generate thousands of possible musical riffs which he has now copyrighted. Thaler is in business and making money.

Hawkins' solutions likely overlap with Dr. Stephen Thaler's patents for neural networks(NN). In particular, Thaler's algorithms inject noise into a proprietary NN system (actually 2 or more NNs conjoined) to generate novel patterns (that is, to _discover_ new patterns). For example, Thaler trained and used such a NN system to generate thousands of possible musical riffs which he has now copyrighted. Thaler is in business and making money.

Check out http://imagination-engines.com/ which is a US company founded by an AI researcher Dr. Stephen Thaler. In summary his systems are composed of paired neural nets in tandem where the first is degraded/excited to produce 'novel ideas' (the 'dreamer') and the second is intended as a 'critic' of the first system's output, or a filter for 'useful' ideas.

In real-life applications, it was used to invent a certain oral-B toothbrush product.

At one time the site's literature announced that 'invention number (CM Creativity Machine) produced invention number 2 (STANNO Self-Training Artifical Neural Object)

Stephen L. Thaler, Ph.D., founded Imagination Engines, Incorporated, a company he created that uses AI. His development of AI centered on pertubing the neural network nodes, and in extreme conditions, destroying the nodes. In the extreme perturbations of continuous node destruction, the AI experienced a "near death experience" whereby it "re-lived" its life and eventually "hallucinated" during the spiral to death. These series of events seem to correlate well in biological systems.

His first patent created the second patent.

Read about it here:

http://www.imagination-engines.com/
http://www.imagination-engines.com/thaler.htm

Stephen L. Thaler, Ph.D., founded Imagination Engines, Incorporated, a company he created that uses AI. His development of AI centered on pertubing the neural network nodes, and in extreme conditions, destroying the nodes. In the extreme perturbations of continuous node destruction, the AI experienced a "near death experience" whereby it "re-lived" its life and eventually "hallucinated" during the spiral to death. These series of events seem to correlate well in biological systems.

His first patent created the second patent.

Read about it here:

http://www.imagination-engines.com/
http://www.imagination-engines.com/thaler.htm

October, 1995, Most Prolific Musician of All Time. A Creativity Machine, trained on top-ten melodies over the preceding 30 years proceeds to invent 11,000 new musical hooks that are promptly copyrighted. Interestingly enough human musical artists disdain the copyright, protesting that "only human minds can conceive music!" Thaler holds back on a million song database generated via Creativity Machine in view of the spirited response from human artists. ...On the flip side, computational musicians don't get it either, feeling that they can do the same thing, spending months or years concocting new works of art via computer. What they don't realize was that only a few hours had been spent by Thaler in translating sheet music to a representation more conducive to neural nets. Beyond that, Creativity Machine function was spontaneous and voluminous.

October, 1995, Most Prolific Musician of All Time. A Creativity Machine, trained on top-ten melodies over the preceding 30 years proceeds to invent 11,000 new musical hooks that are promptly copyrighted. Interestingly enough human musical artists disdain the copyright, protesting that "only human minds can conceive music!" Thaler holds back on a million song database generated via Creativity Machine in view of the spirited response from human artists. ...On the flip side, computational musicians don't get it either, feeling that they can do the same thing, spending months or years concocting new works of art via computer. What they don't realize was that only a few hours had been spent by Thaler in translating sheet music to a representation more conducive to neural nets. Beyond that, Creativity Machine function was spontaneous and voluminous.

This runs the risk of drifting even more offtopic, but what the hell.
Some researchers have found that a neural net can indeed, in some circumstances dream.
Basically, you train it up to recognise faces or tin cans on a production line or whatever, and then you disconnect its inputs. This is equivilent to a human going to sleep. Then the middle layers of the network will drift from state to state, lingering for a time on the various memories it has stored as well as random stuff. This can be read out and displayed by the controling computer or program.
See this link for some more info.