Transition Table Networks Symbolic Signal Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional artificial neural networks face limitations in processing capacity and capabilities, necessitating the development of more sophisticated models that go beyond the neuron-based network structure.
Innovation Solution
Transition Table Networks (TTNs) transmit signals as sets of alphanumeric characters instead of numerical values, using transition tables to determine output strings, allowing for more compact and efficient data encoding and processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional neuron-based neural networks are used, then the network structure is simple and easy to implement, but the processing capacity and capabilities are limited
Solution Approach 1:
The patent changes the fundamental parameters of neural network operation by replacing continuous numerical weight adjustments with discrete symbolic transitions. Transition tables use categorical states and symbolic rules instead of floating-point weights, fundamentally altering how information is processed and transformed through the network, thereby increasing processing capacity while maintaining structural simplicity
Solution Approach 2:
The patent substitutes the mathematical calculus-based mechanism of traditional neural networks with a symbolic rule-based system. Instead of using continuous mathematical functions and gradient descent, the invention employs discrete transition tables with symbolic patterns and rules, replacing the mechanical/mathematical system with a logical-symbolic system that achieves higher processing capabilities
2Productivity
If more hardware is added to increase processing capacity, then processing power improves, but hardware requirements and costs increase
Solution Approach 1:
By changing from continuous numerical parameters to discrete symbolic parameters, the patent enables more complex processing to be achieved with the same hardware resources. The transition table structure allows for richer information representation and processing logic without requiring proportional increases in hardware capacity
Solution Approach 2:
The patent creates a composite computational approach by combining symbolic patterns, transition rules, and categorical states into a unified processing framework. This composite structure enables higher processing density by integrating multiple levels of abstraction (patterns, transitions, rules) within the same computational units
3Productivity
If numerical values are transmitted between nodes, then the network can perform mathematical operations, but data encoding efficiency is reduced
Solution Approach 1:
The patent extracts the essential functional capability of neural networks (information transformation) while removing the inefficient numerical representation layer. By taking out the continuous numerical values and replacing them with discrete symbolic representations, the system achieves more efficient data encoding while preserving the core information processing function
Solution Approach 2:
Instead of representing information as numerical values that require mathematical operations, the patent inverts the approach by using symbolic representations that directly encode meaning. This inversion allows data to be processed in its encoded form without requiring conversion to and from numerical representations, improving encoding efficiency
Data Source
AI summary
A system and method for implementing and training network-based artificial intelligence functions is provided, wherein the signals within an artificial neural network are encoded as strings of symbols such as alphanumeric characters, rather than as numerical weighting values, and accordingly propagated by table functions rather than by summation or sigmoid functions used to handle weighting values. Further, a method is provided for training networks through an evolutionary winnowing and cross-hybridizing process. This different approach to artificial networking is inspired by the biology of genetics, combined with that of neurons, and is anticipated to improve the efficiency and nuance of AI modeling.