Issued U.S. Patent:
| U.S. 6,889,216 | Physical Neural Network Design Incorporating Nanotechnology | [pdf] |
| U.S. 6,995,649 | Variable Resistor Apparatus Formed Utilizing Nanotechnology |
[pdf] |
| U.S. 7,028,017 | Temporal Summation Device Utilizing Nanotechnology |
[pdf] |
| U.S. 7,039,619 | Utilized Nanotechnology Apparatus Using A Neural Network, A Solution and A Connection Gap. | [pdf] |
Pending U.S. Patents
| 20050015351 | ||
| 20040193558 |
Adaptive Neural Network Utilizing Nanotechnology-based Components The concept of using pre- and post-synaptic pulse generation as an asynchronous plasticity mechanism. The general idea is to shape the pre-and post-synaptic pulses, thereby altering how connections are modified via voltage-gradient and frequency dependencies. |
[pdf] |
| 20040162796 | Application of Hebbian and anti-Hebbian Learning to Nanotechnology-based Physical Neural Networks
The concept of using both voltage gradient and frequency dependencies as a mechanism for Hebbian and Anti-Hebbian Plasticity in a Knowm™connection. Mr. Nugent has since used this type of plasticity to design self-repairing pattern recognition and universal logic devices. |
[pdf] |
| 20040153426 | Physical Neural Network Liquid State Machine Utilizing Nanotechnology
The application of Wolfgang Maass's concept of a liquid-state machine as implemented in a Knowm network. The concept is a mechanism for temporal and non-linear separation of signal, allowing a relatively simple mechanism for movement prediction. Mr. Nugent was attracted to this network structure for its relative simplicity and potential computation power. |
[pdf] |
| 20040039717 |
High-Density Synapse Chip using Nanoparticles This patent includes the concept of using two adjacent wafers with perpendicular electrodes as a mechanism for 3-D synapse formation. The process could allow for the stacking of a multitude of independent wafers, which would lead to incredibly dense neural structures. As the connections self-assemble, vertical inter-wafer connections become feasible with current technology. Also of interest is the concept of tailoring synaptic strength by controlling the electrode gap over-lap. |
[pdf] |
| 20030236760 |
Multi-layer Training in a Physical Neural Network formed utilizing Nanotechnology Alex's first attempt at a system-level integration of Knowm™ synapses into a network structure. The design is based on the back-propagation algorithm. The claims cover the general concept of applying a training wave, or a method of sequentially allowing signals within a network to pass from one layer to another. More additions to the general Knowm™ concept, such as the solution comprising a liquid crystal to aid in particle alignment, have been made. |
[pdf] |
|
Plasticity-Induced Self-Organizing Nanotechnology for the Extraction of Independent Components from a Data Stream. The discovery of the flip-lock cycle, a powerful and simple mechanism for the implementation of the AHAH plasticity rule in a Knowm™network. Also of note is the relaxation on uni-directional conducting connections and the adoption of -various inhibitory/excitatory electrode configurations. |
contact |
|
| Universal Logic Gate utilizing Nanotechnology
The application of the flip-lock cycle to small networks in combination with simple CMOS core circuitry. The result is a CMOS circuit composed of about 30 transistors, as well as a small Knowm synapse matrix above the core circuitry capable of being configured to implement any logic function. The circuit assembles and repairs its connection as it processes information. |
contact us |
|
|
Methodology for the Configuration and Repair of Unreliable Switching Elements The application of the flip-lock cycle, as well as AHAH plasticity to molecular switching elements, crossbar bi-stable switches, and meta-stable switches in general. The methodology developed for the universal logic gate is applied to other competing nanoscale connection elements, taking Knowm™ self-assembling technology out of a liquid suspension and into the general nanoelectronics arena. |
contact us |
|
|
General Purpose Dielectrophoretic Immuno Assay Technique A low-cost immuno assay technique that utilizes positive dielectrophoresis to attract latex beads to a connection gap, where antigen is bound to the beads and/or chip surface. By monitoring the electrode capacitance while negative dielectrophoresis is used to repel the beads, a measurement of antigen bond force, and therefore concentration, can be made by a simple, automated, electronic system. The technique dramatically reduces the current cost of such measurements and is extremely simple, both in fabrication and operation. |
contact us |
|
Fractal Memory and Computational Methods and Systems Based on Nanotechnology Fractal Memory™ is a robust and infinitely scalable architecture for distributed pattern recognition. Similar to content-addressable memory, Knowm Fractal Memory™ is specifically designed for universal pattern recognition and classification tasks and offers a number of features not found in any current technology. The architecture can perform probabilistic best-match pattern recognition and arbitrary classification of an unlimited pattern template database, as well as providing a quantitative measure of pattern recognition certainty, all in one clock cycle. In addition, the architecture provides on-chip learning, adaptation and repair and can be taught to recognize new patterns without taking the device off-line. |
contact us |