Self-Learning Network Training via Interpolated Input Sets
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional modeling platforms struggle when necessary inputs are unknown, but a target output is known, as analysts must manually estimate and adjust component values, leading to inefficiencies and potential ripple effects, without leveraging the decision logic or patterns in interpolation engines for machine learning.
Innovation Solution
Systems and methods for training self-learning networks using interpolated input sets based on a target output, where the network accesses and learns from data generated by interpolation platforms, replicating decision-making patterns and logic to automate the process of producing known outputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If analysts manually estimate and adjust component values to achieve target outputs, then the target output can be achieved, but the process is time-consuming and prone to ripple effects requiring multiple iterations
Solution Approach 1:
The patent creates virtual copies of the interpolation engine's decision logic by training machine learning models (neural networks, fuzzy logic systems) to replicate the interpolation engine's input-output behavior. These copied models can rapidly generate interpolated inputs without manual iteration, achieving the same target outputs while eliminating time-consuming manual adjustments.
Solution Approach 2:
The patent performs preliminary training of machine learning models using historical interpolation data before actual use. This preliminary action pre-configures the models with decision logic patterns, enabling them to rapidly generate accurate interpolated inputs without requiring manual estimation and adjustment during actual operations.
2Measurement precision
If analysts manually perform interpolation to determine input distributions, then the target output can be achieved, but the process requires significant analyst involvement and iterative adjustments
Solution Approach 1:
The patent enables the system to perform interpolation automatically through trained machine learning models that self-service by generating interpolated inputs without human intervention. The models learn decision logic patterns from historical data and autonomously produce accurate interpolated inputs, eliminating the need for analyst involvement in the actual interpolation process.
Solution Approach 2:
The patent replaces the mechanical manual process of analyst estimation and adjustment with automated machine learning systems. Neural networks and fuzzy logic systems substitute for human analysts, using learned patterns to automatically generate interpolated inputs with high accuracy without requiring manual iteration or adjustment.
3Loss of information
If the interpolation engine generates interpolated inputs based on target outputs, then the decision logic can be captured, but this valuable pattern is not exploited for machine learning
Solution Approach 1:
The patent creates copies of the interpolation engine's decision logic by training machine learning models to replicate its input-output behavior. These copied models capture the decision logic patterns embedded in the interpolation engine and can be deployed independently to generate interpolated inputs, preserving the valuable decision logic while enabling broader application.
Solution Approach 2:
The patent performs preliminary training of machine learning models using historical interpolation data that encapsulates the decision logic. This preliminary action extracts and stores decision logic patterns in the trained models, making them available for future use in various scenarios without requiring access to the original interpolation engine or manual analysis of decision logic.
Data Source
AI summary
Embodiments relate to systems and methods for training a self-learning network using interpolated input sets based on a target output. A database management system can store sets of operational data, such as financial, medical, climate or other information. A user can input or access a set of target data, representing an output which a user wishes to be generated from an interpolated set of input data. The interpolation engine can generate a conformal interpolation function and input sets that map to the set of target output data. After interpolation, the interpolation engine can transmit the interpolated inputs, along with the set of target output data and other information, to a self-learning network such as a neural or fuzzy logic network. The self-learning network can be trained to converge to the target output based on the interpolated input results as generated by the interpolation engine, thus reproducing the desired interpolation function.