Feature Vector Rearrangement for Human State Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional machine learning requires a large number of labeled data samples and a small number of dimensions for accurate inference of human internal states from high-dimensional biological data, making it challenging to achieve accurate results with limited data sets.
Innovation Solution
A computer system that rearranges feature vectors in a feature space based on defined forces, including attractions and repulsions, to minimize potential energy, allowing for more accurate inference from a limited data set by rearranging feature vectors to reduce potential energy and improve model training and prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional machine learning is used to infer human internal states from biological data, then the inference accuracy can be improved, but the requirement for large amounts of labeled data and small dimensionality increases
Solution Approach 1:
The patent transforms the high-dimensional biological data by changing the parameter representation through force-based rearrangement in feature space. By defining attraction and repulsion forces between feature vectors and rearranging them to minimize potential energy, the method effectively reduces the dimensional complexity while preserving the essential information needed for accurate inference, thereby reducing the amount of labeled data required.
Solution Approach 2:
The patent introduces an intermediary process of force-based feature vector rearrangement between the raw high-dimensional biological data and the classification/inference stage. This intermediary transformation reorganizes the feature space by applying attraction and repulsion forces, creating a more efficient representation that reduces the data quantity requirement while maintaining inference accuracy.
2Measurement precision
If traditional machine learning is used to infer human internal states from biological data, then the inference accuracy can be improved, but the data dimensionality increases
Solution Approach 1:
The patent applies parameter changes by transforming the high-dimensional feature space through force-based rearrangement. By defining potential energy based on attraction and repulsion forces between feature vectors and minimizing this energy, the method reorganizes the data representation to reduce effective dimensionality while preserving the discriminative information necessary for accurate inference.
Solution Approach 2:
The patent addresses dimensionality by introducing a new organizational dimension in the feature space through force-based rearrangement. Instead of directly reducing dimensions, it reorganizes feature vectors along force-directed dimensions defined by attraction and repulsion relationships, creating a more compact and efficient feature representation that reduces complexity while maintaining accuracy.
3Measurement precision
If feature vectors are rearranged to reduce potential energy, then the accuracy of inferring labels from limited data increases, but the computational complexity of the processing increases
Solution Approach 1:
The patent implements self-service by allowing the feature vectors to automatically rearrange themselves through the force-based optimization process. The attraction and repulsion forces create a self-organizing mechanism where feature vectors autonomously find their optimal positions in the feature space by minimizing potential energy, reducing the need for complex external computational intervention while improving inference accuracy.
Solution Approach 2:
The patent applies preliminary action by pre-rearranging the feature vectors through force-based optimization before the actual inference process. This preliminary reorganization of the feature space, based on attraction and repulsion forces, prepares the data in an optimal state that enhances subsequent inference accuracy from limited data while the computational cost is incurred in advance.
Data Source
AI summary
A computer system includes a storage device and a processor configured to operate in accordance with command codes stored in the storage device. The processor places a plurality of feature vectors including feature vectors each provided with a label in a feature space. The processor determines forces among the plurality of feature vectors determined from repulsions and attractions between feature vectors. A repulsion is defined to be larger when a distance between the feature vectors is shorter, and an attraction is defined to be larger when the feature vectors have a predetermined relationship. The processor rearranges the plurality of feature vectors to reduce potential energy of the feature space determined by the forces among the plurality of feature vectors. The processor outputs at least a part of the plurality of rearranged feature vectors as data to be used for data analysis.