18071231. NEURAL NETWORK WORD CLUSTERING SYSTEM simplified abstract (SAP SE)
Contents
- 1 NEURAL NETWORK WORD CLUSTERING SYSTEM
- 1.1 Organization Name
- 1.2 Inventor(s)
- 1.3 NEURAL NETWORK WORD CLUSTERING SYSTEM - A simplified explanation of the abstract
- 1.4 Simplified Explanation
- 1.5 Potential Applications
- 1.6 Problems Solved
- 1.7 Benefits
- 1.8 Potential Commercial Applications
- 1.9 Possible Prior Art
- 1.10 Unanswered Questions
- 1.11 Original Abstract Submitted
NEURAL NETWORK WORD CLUSTERING SYSTEM
Organization Name
Inventor(s)
MAREK Polewczyk of Berlin (DE)
Marco Spinaci of Versailles (FR)
NEURAL NETWORK WORD CLUSTERING SYSTEM - A simplified explanation of the abstract
This abstract first appeared for US patent application 18071231 titled 'NEURAL NETWORK WORD CLUSTERING SYSTEM
Simplified Explanation
The abstract describes a neural network clustering system that detects bounding boxes, generates an adjacency matrix, and clusters words into different lines on a document.
- Detects bounding boxes and identifies coordinates for each
- Generates an adjacency matrix based on key and query matrices
- Clusters words into different lines on a document
- Arranges words on a second document based on clusters
- Provides the second document for display
Potential Applications
This technology could be applied in document organization, data visualization, and information retrieval systems.
Problems Solved
This technology solves the problem of efficiently organizing and displaying large amounts of text data in a visually appealing and structured manner.
Benefits
The benefits of this technology include improved document readability, easier information retrieval, and enhanced data organization.
Potential Commercial Applications
- Document management software
- Data visualization tools
- Information retrieval systems
Possible Prior Art
One possible prior art for this technology could be traditional document clustering algorithms used in natural language processing systems.
Unanswered Questions
How does this technology handle overlapping bounding boxes?
The abstract does not mention how the system deals with overlapping bounding boxes and if it affects the clustering process.
What is the computational complexity of this clustering system?
The abstract does not provide information on the computational resources required to run this neural network clustering system.
Original Abstract Submitted
Various embodiments for a neural network clustering system are described herein. An embodiment operates by detecting a plurality of bounding boxes and identifying coordinates for each of the bounding boxes. An adjacency matrix is generated based on combining a key matrix and a query matrix. The plurality of words are clustered into a plurality of clusters, each cluster corresponding to a different line on the first document. A second document is generated in which the plurality of words corresponding to a respective cluster of the plurality of clusters is arranged on a same line on the second document. The second document is provided for display.
