Multi-Model Classification for Faster VAE Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing classification methods using Variational Autoencoders (VAEs) require building a VAE for each class, leading to high throughput demands and prolonged operation times.
Innovation Solution
Implement a method using multiple machine learning models, each configured to classify into different classes, allowing for parallel processing and reduced operation time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a VAE is built for every class to be judged, then classification accuracy is improved, but the amount of throughput increases and operation time is prolonged
Solution Approach 1:
The patent segments the classification task by dividing the plurality of classes into multiple groups, where each group is handled by a separate VAE model. This allows parallel processing of different class groups, improving throughput while maintaining classification accuracy through specialized models for each group.
Solution Approach 2:
The patent creates a universal classification system where multiple VAE models share a common architecture and can collectively handle all classes. Each VAE is trained on a subset of classes but the overall system provides comprehensive classification coverage, enabling both accuracy and efficiency.
2Measurement precision
If a VAE is built for every class to be judged, then classification accuracy is improved, but operation time is prolonged
Solution Approach 1:
The patent segments the classification task by dividing the plurality of classes into multiple groups, where each group is handled by a separate VAE model. This allows parallel processing of different class groups, improving throughput while maintaining classification accuracy through specialized models for each group.
Solution Approach 2:
The patent performs preliminary grouping of classes and pre-trains multiple VAE models on their respective class groups before actual classification. This preparation enables faster inference time during operation, as the models are already optimized for their specific class groups and can process data in parallel.
Data Source
AI summary
A method of making a single processor or a plurality of processors perform classification processing of classification target data using a machine learning model includes the steps of (a) preparing N machine learning models in a memory assuming N as an integer no smaller than 2, and (b) performing the classification processing of the classification target data using the N machine learning models. Each of the N machine learning models is configured so as to classify input data into any of a plurality of classes with learning using training data, and is configured so as to have at least one class different from a class of another of the N machine learning models.