Holographic Wave Phase Storage in a Novel Panassociative Circuit

by Steven R. Grimm srgrimm@fuse.net and Ronald C. Blue ronblue@enter.net, JCIS 2003, 7th Joint Conference on Information Sciences, September 26-30, 2003, Embassy Suites Hotel and Conference Center, Cary, North Carolina, USA

Keywords: Holographic, quantum, attractors, panassociative, correlational holographic opponent processing, mind, brain function.

Abstract: Using a novel approach to store and retrieve sensor data, we will demonstrate how hardware modules incorporating holographic storage principles can be linked together to provide particular functions of the brain. These modules will illustrate characteristics of interference memory, local and global stabilization and dynamic entangled states.

The Panassociative circuit is an isotropic matrix of passive NPN transistors with associated diodes for directing signal path. For further details on passive transistors being used as computation devices, refer to: http://www.enter.net/~ronblue/index2.htm; http://biobotics.150m.com; and http://www.enticypress.com

In a mobile machine comprising of audio, visual and obstacle detection sensors, the inputs of each sensor is brought into the circuit in a matrix configuration directly connecting to the passive transistor/diode structure each occupying a local position in the matrix. There is no signal conditioning or amplification of sensor data before connection to the transistors and each signal modified by the sensor is initially derived from the ground side of the power supply, thereby introducing an active negative into the circuit whereas the positive supply acts as a stable reference source introduced onto the collector side of the transistor.

The sensor input values are connected to the emitter side of the transistor while the base acts as the resultant output. The operative condition within the transistors is opponent processing whereby fields of opposing polarity produce macroscopic standing waves further modulated by sensory input. The positive condition introduced to the common collector sides of the transistors are modulated by a sine wave signal operating at approximately 6 to 7 kHz. This sine wave produces a carrier type wave formation through the circuit resulting in a consistent wave phase at each transistor location. The sensor input, due to opponent processing, modulates the sine wave field set up in one half of the transistor inducing a computed output of both opposing fields. This process occurs in each transistor location and contributes it's output to all other transistors. The overall condition of the modulated wave signals are of a global nature since each part of each modulated singular wave contributes to the whole while on a local level a higher amplitude signal from each sensor effects that particular transistor and it's closest neighbor.

These circuits, configured as separate modules can represent particular functions of the brain depending on the interconnections and the direct influence of it's sensors, whether they are external to the machine or from internal states. Examples of these functions would include short term memory, long term memory, attention, motivation, learning and postulating of concepts, to name a few.

Theoretically, enough of these modules would be able to facilitate LTM as superpositional compressed states of association. LTM would be "called up" when a resonance between STM and LTM is established due to either sensory status or internal associations contributed to "thought". This agrees with Steve Grossberg's theory of Adaptive Resonance (ART) and demonstrates the top-down, bottom-up signal correlation of resonance.

In the present model, it's more efficient to demonstrate the Pan-circuits as STM as well as how macro-quantum fields can be generated and sustained providing "awareness of the present" in relation to internal and external machine states.

Recent research has provided new information (i.e. P.T. Chopping:The Holographic Brain,10/2000, http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/M/Memory.html, and “Memory in Aplasia” at <http://users.rcn.com/jkimball.ma.ultranet/BiologyPages/

M/Memory.html> ) that in biological LTM processing, protein synthesis is indeed a part of forming long term facilitation at the synaptic level and in forming new synaptic connections. This information has led me back to investigating the use of software in artificial neural networks as LTM.  The Hybrid neural net that I have designed works in concert with STM and it's inherent wave properties. The first characteristic one will notice is that there is no on-line, off line learning time; the "learn" feature is always on and continuously adaptive. The "weights" to be adjusted are to be viewed as the synaptic strengths generated as learning progresses through a type of Hebbian process (D.O. Hebb); i.e. protein synthesis. The outputs of a standard ANN, particularly ones that employ backpropagation are generally static in nature and provide a one to one correspondence to the input: (input=>function=>output). The Hybrid net provides a continuous output necessarily, because of the way it feeds back into the Pan networks. This is the first established link between STM and LTM resonance. However, it must be kept in mind that there are as many "soft" networks as there are hardware modules. The introduction of the response of a LTM into the STM modules changes the overall activity of the wave functions on the global level and obviously the local level. These are reworked memories and associations.

A compression of sorts results from this resonance of STM to LTM and back and facilitates an oscillation "field" within the hardware modules. This becomes a global field and helps in binding the modules at the level of the wave phase and further promotes phase lock at resonance. The "compression" is manifested through translation of hardware signal frequencies to software data samples; as in sampling any other type of waveform. At particular intervals, the sampling software reads the holographic signals from the Pan circuits and are transferred to their particular software network. The networks are separate in function but architecturally connected. Input node X(1) may have an affect on weight node W(1,2) but also affect weight node W(1,3,2) in another network. The data capture from the sampling software is performed by analog to digital conversion at a high rate of speed. More importantly is that the data being sampled is not that of standard DC voltage but of the superposed frequencies from the Pan circuits. These frequencies represent the spectral information encoded in the phases of the input waveforms. A transform of the superposed frequencies is performed due to the resolution of the sampling rate and the resolution of bit conversion in the A/D converter. This results in generally a lower, "edited" version of the original signal.

An algebraic conversion takes place within the Hybrid-net and will produce a potential output to a digital to analog (D/A) converter which places the associated signal (now a frequency) back into the hardware Pan circuit network, but only if the output is relevant to the input. This establishes an auto-correlation between STM and LTM and provides a linear "tracking" of temporal states from one event to another. It is also in this process that a level of awareness and perhaps even consciousness may be realized. According
to the literature pertaining to this subject (Mitchell, Pribram, Wolf, Pietsch, Gupta, Marcer, and Chopping) as well as others (Penrose and Hammeroff), indicate that a process rather than a "place" is responsible in bringing forth a conscious state. This state or series of states usually accompany a "transform" of information from it's stored state to an active one. Whether this is the collapse of a wave function in micro-tubels, phase conjugant adaptive resonance (pcar), reconstructing a holonomic wavefront or emergent conscious qualities from extreme parallel networks; they all tend to point in the direction of an energy/information transformation of states.

In the Pan circuit, the information transform takes place as an interaction of correlated informational energies of opposite polarities at the atomic level within a neutral chamber (and field) of a NPN transistor. But it also takes place in subsequent levels of interaction as one builds upon the other.

Evan Harris Walker modified the Schrodinger equation to make potential fields physically effective. The Schrodinger equation psi has an imaginary conjugate psi* which is not physically effective in the standard model. He has suggested that consciousness may be associated with psi* which is essential to his equations. He states that psi and psi* can not exist without one another as one completes the other to yield a real physical result. Each are relative potentials until they come together and then produce an observable output. By analog, that is the same process that occurs within the NPN transistor. However, we are using the electrons and positive holes on either side of a transistor substrate to perform a neutral computation of those opponents. Standard measurement of the output of the transistor (base region) will indicate a very strong noise component. That's because the measurement is looking for the positive side of a signal, since the measurement is normally taken in reference to the negative terminal of the power supply. In this case, the true output is masked an unavailable. It only shows up when the reference method is reversed and the positive terminal is used as the stable reference. It will then provide a measurement of the variable negative output, yet, both sides of the signal are simultaneously present. This is very much like the Psi* equation of Shrodinger where we would encounter such concepts of "zero-space" and negative space. P.Peitsch of ShuffleBrain calls this concept "Active Zero".

The incoming frequency wave generated by the sensory circuits are time variant. They occur in real linear time. This is also the next level in which signal transformation occurs. Taken at the level of the waveform, phase information becomes diffused and distributed throughout the Pan circuit creating a holographic superposition of information about the incoming data. This is further fused with the wave transformation frequencies generated by the D/A conversion via association.  We now have an oscillatory effect throughout the system that will maintain itself, be self stabilizing and self organizing. Output circuits that translate components of the holographic system are used for motor control as well as feedback to update internal states of the machine.