Microsoft patents method for grouping similar text into visual clusters
Microsoft describes a system that turns text records into mathematical vectors to measure how similar they are. It then builds a graph where similar items are connected, allowing the software to identify groups of related content and display them visually.
In the filing’s own words
Example solutions for clustering data include: encoding a plurality of source records into a plurality of vectors, each source record containing source terms being encoded as one vector of the plurality of vectors; computing a similarity score for each unique pair of vectors of the plurality of vectors; constructing a similarity graph by: adding a node to the similarity graph for each vector; and adding an edge between each pair of nodes in which the similarity score for the associated pair of vectors exceeds a first similarity threshold; identifying a cluster of nodes within the similarity graph, the cluster of nodes representing a disconnected subgraph within the similarity graph; and generating a graphical representation of the one or more clusters on a display device.
What exactly they claim (claim 1)
1. A clustering system comprising: a processor; and a computer-readable medium storing instructions that are operative upon execution by the processor to: encode a source record into a first vector, the source record containing source terms being encoded by the first vector; compute a similarity score for a first pair of vectors including the first vector and a second vector, the similarity score being a value representing how similar the first vector is to the second vector; construct a similarity graph by: adding a first node and a second node to the similarity graph for the first and second vectors; and adding an edge between the first node and the second node based on the similarity score exceeding a first similarity threshold; identify a first cluster of nodes within the similarity graph, the first cluster of nodes representing a disconnected subgraph within the similarity graph, the first cluster of nodes including the first and second nodes; and generate a graphical representation of the first cluster of nodes on a display device.
Facts
- Company
- Microsoft (Microsoft Technology Licensing, LLC)
- Official title
- Uncovering patterns in text through clustering
- Patent number
- US12743826B2
- Application number
- 18644076
- Filed
- Apr 23, 2024
- Granted
- Sep 22, 2026
- Inventors
- 4
- Claims
- 20
Classified under
G06T 11/26 Physics & computingG06F 16/35 Physics & computingG06F 16/9024 Physics & computingG06F 21/6254 Physics & computingMore from Microsoft in AI & software
- Microsoft patents method to check AI answers against verified facts Oct 20, 2026
- Microsoft patents AI that presents PowerPoint slides for you Sep 3, 2026
- Microsoft patents method to check if LLM answers are based on real data Sep 10, 2026
- Microsoft patents system to blend multiple AI agents into one chat Oct 1, 2026
- Microsoft patents system that answers questions using your screen history Oct 15, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.