Clustering Apparatus Using Dual Trained Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing clustering apparatuses face challenges in accurately clustering target data due to the presence of harmful data such as unidentifiable and noisy data, which can hinder correct clustering, reduce performance, and obscure cluster boundaries.
Innovation Solution
The proposed clustering apparatus utilizes two trained models to process target data, where the first model generates a first feature vector and the second model generates a second feature vector. These vectors are then integrated to improve clustering performance and visibility, with the second model serving as an auxiliary to enhance the first model's output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single trained model is used for clustering target data, then the device complexity is low, but the clustering performance and accuracy deteriorate due to harmful data
Solution Approach 1:
The patent combines multiple trained models (first trained model and second trained model) to process target data together. The first model generates first feature vectors while the second model generates second feature vectors, and these are integrated to produce clustering results. This merging of multiple models improves clustering performance and accuracy by compensating for the weaknesses of individual models when dealing with harmful data.
Solution Approach 2:
The patent introduces an integration mechanism that acts as an intermediary between the two trained models and the final clustering output. This integration unit combines the feature vectors from both models and coordinates their outputs to generate the final clustering results, thereby improving reliability while managing the complexity through structured integration.
2Measurement precision
If multiple trained models are used to process target data, then the clustering accuracy improves, but the calculation time and processing complexity increase
Solution Approach 1:
The patent segments the processing task by dividing it into two parallel model processing streams. The first trained model processes target data to generate first feature vectors, while the second trained model simultaneously processes the same data to generate second feature vectors. This segmentation allows for parallel computation, improving accuracy while managing calculation time through efficient task division.
Solution Approach 2:
The patent uses two models instead of one, performing slightly more processing than a single model would require. This partial excess action (using an additional model) improves clustering accuracy by providing multiple perspectives on the data, while the integration mechanism ensures the additional processing time is utilized efficiently rather than redundantly.
3Quantity of substance
If harmful data is included in the target data group, then the quantity of data increases, but the clustering performance and visibility of results deteriorate
Solution Approach 1:
The patent converts the harmful effect of diverse data (including potentially harmful data) into a benefit by using multiple trained models that can process and interpret the same data from different perspectives. The integration of multiple models' outputs helps identify and mitigate the impact of harmful data while utilizing the full data set, thereby maintaining data quantity while improving clustering performance.
Data Source
AI summary
A clustering apparatus includes processing circuitry. The processing circuitry is configured to: acquire target data, a first trained model adapted to receive input of the target data and output a first feature vector, and a second trained model adapted to receive input of the target data and output a second feature vector; calculate the first feature vector using the first trained model and the target data; calculate the second feature vector using the second trained model and the target data; calculate a second cluster by dividing the second feature vector; and integrate the first feature vector with the second cluster.