Building synthetic brains capable of human level discovery and invention...

Summary - An artificial neural network that has been trained on some body of knowledge and then perturbed in a specially prescribed way tends to activate into concepts and/or strategies (e.g., new ideas) derived from that original body of knowledge. These transiently perturbed networks are called 'imagination engines' or 'imagitrons'. If another computational agent, such as a traditional rule-based algorithm or, even better, another trained neural network is allowed to filter for the very best of these emerging ideas, we arrive at an extremely valuable neural architecture, the patented Creativity Machine. Optional feedback connections between this latter computational agent and the imagination engine assure swift convergence toward useful ideas or strategies. ..This new AI paradigm is vastly more powerful than genetic algorithms (GA), efficiently generating new concepts on mere desktop computers rather than on the computational clusters required of GAs.

Details - Ordinary neural networks excel at learning from raw data, cumulatively learning to associate one pattern with another, as when raw sensory inputs from our five senses activate mental images or feelings. Note, however, that such a direct link between the external world, and our own internal mental life is only a small part of brain activity. To produce a more faithful emulation of human cognition, some mechanism must be established that provides an internal genesis of thoughts and ideas that draws upon cumulative experience, rather than what our senses are telling us at the moment.

Creativity Machines represent a new kind of neural network paradigm that is capable of generating rather than just associating patterns. They are based upon what we believe to be a significant scientific discovery: that a neural network exposed to any conceptual space and then internally irritated, in a specially prescribed way, tends to generate coherent ideas derived from its absorbed wisdom. Another neural network, trained to filter out the very best of these notions, patrols the former net's outputs, accumulating useful concepts (i.e., new drug or automobile designs) or using these output patterns in real time to devise strategies (i.e., robotics and control systems).

Perhaps the most appropriate benchmark for Creativity Machines is the highly popularized genetic algorithm (GA) wherein mathematically simulated 'genes' randomly combine and mutate to produce new potential offspring that in turn represent new concepts, designs, or strategies. The problem with such systems is that they work on geologic time scales relative to the inherent processing speed of  digital computers. Furthermore, GAs fail to deliver solutions as the dimensionality of the problem increases and the algorithm's generative process leads to combinatorial explosion and nonsensical offspring. 

Problem dimensionality is not an issue for Creativity Machines. In a nutshell, this advantage stems from the fact that this neural architecture emulates the thalamo-cortical loop of the brain (i.e., the seat of intelligence and consciousness) rather than blind, and excruciatingly slow processes of mutation and natural selection. Furthermore, the actual time to build a Creativity Machine is negligible to genetic programs, since the CM is a self-organizing system. GA's, in contrast are not, and require that human experts hardwire myriad constraint relationships into their human-originated computer codes. In other words, genetic algorithms must be written by human programmers, whereas Creativity Machines build themselves!

The benefit to you, as an IEI customer, is that a specially tailored Creativity Machine can generate results at typically a tenth of the cost of any other AI paradigm perceived as competitive. Furthermore, it can deal with complex problems that the other approaches can't even touch.

The movie below illustrates a striking example of a Creativity Machine project. A million input-output STANNO (IEI's totally autonomously training neural network) has been exposed to bitmaps of 12 different faces over a period of 1 minute. Running freely, this Creativity Machine produces a wide gamut of new, potential faces, generalized from the STANNO's training set, and all distinct from those exemplary faces. Compare the invented faces within the movie sequence with the the training exemplars surrounding it.

In contrast, a genetic algorithm would require months or even years to build as programmers capture all of the implicit constraints that qualify a pattern of pixels as a face. When finally run, the GA would inevitably require years of run time, on the fastest computers in the world to build a single coherent face. Results such as these are exactly why the Creativity Machine has been heralded by top NASA officials as AI's best bet and the primary tool for building the AI predicted by Kurzweil and others in 30 years, now!

Yam, Philip (1993). "Daisy, Daisy: do computers have near-death experience?", Scientific American, May 1993.

Yam, Philip(1995). "As They Lay Dying ... Near the end, artificial neural networks become creative", Scientific American, May, 1995.

Thaler, S. L. (1996). Neural Networks That Create and Discover, PC AI, May/June 1996.

Holmes, R. (1996). The Creativity Machine, New Scientist, 20 January 1996.

Thaler, S. L. (1997). US Patent 5,659,666, "Device for the autonomous generation of useful information", Issued August 19, 1997.

Thaler, S. L. (1997). US Patent 6,018,727, "Device for the autonomous generation of useful information", Issued January 25, 2000.

Thaler, S. L. (2002). US Patent 6,356,884, "System for the Autonomous Generation of Useful Information", Issued March 12, 2002.