Dynamic Composite Data Structures Using Feature Envelopes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current modeling techniques often produce unrealistic or unnatural variations of data due to the high dimensionality of analyzed datasets, which complicates the preservation of natural correlations between parameters, especially in complex modifications, and are not suitable for low-footprint environments.
Innovation Solution
The method involves building and training function mappers, such as multilayer perceptron neural networks, to map extracted feature envelopes to analyzed datasets, allowing for the combination and modification of data segments while preserving correlations, and applying magnitude-dependent weighting functions for training and synthesis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional modeling techniques are used to generate data variations, then the process is simple, but the generated variations are unrealistic and unnatural
Solution Approach 1:
The patent segments the original data into multiple data segments and extracts feature envelopes from each segment. This segmentation allows the system to manage complexity by breaking down the overall modeling task into smaller, more manageable parts while preserving the natural characteristics of each segment for more realistic synthesis.
Solution Approach 2:
The patent transforms the high-dimensional analyzed dataset into a lower-dimensional feature envelope representation. This dimensionality reduction simplifies the data structure while maintaining the essential correlations, enabling realistic data generation without requiring complex high-dimensional processing.
2Loss of information
If high dimensionality analysis is performed on datasets, then detailed information is captured, but natural correlations between parameters are lost
Solution Approach 1:
The patent reduces the dimensionality of the analyzed dataset by extracting feature envelopes that capture the essential correlations between parameters. This transformation maintains the natural relationships in the data while working with a more manageable dimensional structure.
Solution Approach 2:
The patent extracts feature envelopes from the high-dimensional analyzed dataset, isolating the key characteristics and correlations needed for realistic data synthesis. This extraction process removes unnecessary complexity while preserving the essential natural correlations.
3Adaptability or versatility
If complex modifications are applied to data, then versatility is improved, but preserving natural correlations becomes difficult
Solution Approach 1:
The patent segments the data into multiple segments with their own feature envelopes, allowing independent modification of each segment while maintaining the natural correlations within each segment. This segmentation enables versatile complex modifications without losing the authentic characteristics of the original data.
Solution Approach 2:
The patent uses dynamically controllable composite data structures that can be adjusted and modified in real-time. This dynamic approach allows for versatile modifications while preserving natural correlations through the use of feature envelopes that capture the essential relationships.
4Ease of manufacture
If traditional data structures are used, then simplicity is maintained, but they are not suitable for low-footprint environments
Solution Approach 1:
The patent reduces data dimensionality by representing complex datasets as composite structures of feature envelopes. This dimensionality reduction makes the data structures more suitable for low-footprint environments by requiring less memory and computational resources while maintaining essential information.
Solution Approach 2:
The patent extracts only the essential feature envelopes from the original data, removing unnecessary complexity and reducing the data footprint. This extraction creates streamlined data structures that are efficient for storage and processing in resource-constrained environments.
Data Source
AI summary
A method of producing dynamic controllable data composites from two or more data segments includes: building or training one or more function mappers to map between one or more extracted feature envelopes sets from the original data and one or more general parametric representations of the data; combining the extracted feature envelopes or the function mappers using two or more audio segments; and feeding the extracted feature envelopes or combined feature envelopes to the function mappers to obtain synthesis parameters to drive a synthesis process.