Binary Classifier Suite for Faster Multi-Class Object Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Generating multi-class or N-way machine learning classifiers is time-consuming, resource-intensive, and complex, often outweighing the benefits of automating data classification tasks.
Innovation Solution
Implement a set of hyper-trained binary classifiers to divide the characteristic range into subranges, with each classifier determining the presence of a specific subrange, and aggregate their outputs to generate a response curve for object identification and characteristic determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a multi-class or N-way machine learning classifier is used to perform comprehensive object classification and characteristic detection, then classification accuracy and automation capability are improved, but training time and computational complexity increase significantly
Solution Approach 1:
The patent divides a complex N-way classification problem into multiple binary classification problems. Instead of training one large multi-class classifier, the system creates multiple binary classifiers, each trained to distinguish between two specific classes or characteristics. This segmentation reduces the training complexity and time for each individual classifier while maintaining the overall classification capability through aggregation of binary classifier outputs.
2Adaptability or versatility
If a multi-class or N-way machine learning classifier is used to perform comprehensive object classification, then classification capability is improved, but computational resources and complexity increase significantly
Solution Approach 1:
The complex N-way classification task is segmented into multiple simpler binary classification tasks. Each binary classifier handles a specific subset of classes or characteristics, reducing the computational burden on each model. The system aggregates results from multiple binary classifiers to achieve comprehensive classification, thereby maintaining high adaptability while reducing individual model complexity.
Solution Approach 2:
Multiple binary classifiers are combined to achieve the functionality of a single N-way classifier. The outputs from individual binary classifiers are aggregated through various methods (such as voting, probability combination, or decision fusion) to produce the final classification result. This merging approach allows the system to leverage the simplicity of binary classifiers while achieving the versatility of multi-class classification.
3Reliability
If a multi-class or N-way machine learning classifier is used to detect multiple characteristics, then detection comprehensiveness is improved, but training data requirements and resource consumption increase
Solution Approach 1:
The patent segments the detection of multiple characteristics into separate binary classification tasks. Each binary classifier is trained on a specific characteristic or class distinction, requiring less training data per classifier compared to training a single N-way classifier on all classes and characteristics simultaneously. This segmentation allows for more efficient use of training data while maintaining comprehensive detection capability.
Data Source
AI summary
Described herein are systems and methods for generating multi-class machine learning classifiers using a plurality of hyper-trained binary classifiers to perform both characteristic detection and object type classification. A range associated with a characteristic of the data sought to be classified may be divided into a plurality of sub-ranges that can be ordered. A plurality of binary machine learning classifiers can be generated, with each binary classifier associated with one of the sub-ranges. Each binary classifier of the plurality of classifiers may be generated by training the classifier to determine if a given input data set contains a characteristic within or outside of the sub-range associated with the binary classifier. Once the binary classifiers have been trained, they can be used to determine what value within the range of the characteristic an input data sample has, and may be used to determine other properties about the input data.