ken365 OP : It seems like this. It's not well understood.
Patent 1: US 12,238,136 B2
Patent Number: US 12,238,136 B2
Title of the Invention: Malware Data Clustering
　　　　　(Malware Data Clustering)
Inventors Harkirat Singh (New York, NY, USA)
　　　　Geoffrey Stowe (San Francisco, CA, USA)
　　　　Stefan Baak (Isleworth, GB)
Matthew Sprague (Palo Alto, CA, USA)
Michael Kross (Palo Alto, CA, USA)
Adam Borochoff (New York, NY, USA)
Parvathy Menon (San Jose, CA, USA)
Michael Harris (Carmel, IN, USA)
Applicant: Palantir Technologies Inc. (Denver, CO, USA)
Application Number: 18/504,392
Filing Date: November 8, 2023
Patent issuance date: February 25, 2025
Related applications:
Continuation application: US 17/658,893
(Filed on April 12, 2022, Patent number US 11,848,760)
International Patent Classification (IPC):
・G06Q 40/00 (Information Processing and Data Analysis)
・G06F 16/23 (Data Management Technology)
USA Patent Classification (CPC):
・H04L 63/145 (related to Security)
・G06F 16/23, G06F 16/244 (Data Clustering and Search)
▷Overview of the invention
This patent relates to technologies for clustering related data such as malware. Specifically, a method is shown for grouping related data using a "Seed" to create clusters.
(1) Steps for cluster generation
・First, obtain the "Seed" of the data.
・Add the Seed to the initial cluster.
・Apply the clustering strategy or rules to add data related to the Seed to that cluster.
⑵ Evaluation and Scoring of Clusters
・Generate "Cluster Score" based on the characteristics of the cluster.
・The Cluster Score is determined based on the attributes of the data.
・Moreover, create a "Cluster Metascore (overall evaluation of the cluster)" using the Cluster Score.
⑶ Application Areas
・Classification and analysis of malware
・Detection of tax fraud
・Detection of "beacon-type" malware (malware that regularly communicates with the outside).
By utilizing this technology, analysts can more efficiently discover patterns in data and improve the accuracy of Cybersecurity and fraud detection.
Patent 2: US 12,235,829 B2
Patent number: US 12,235,829 B2
Title of the invention: Systems and Methods for Generating Interrelated Notional Data
(System and Methods for Generating Interrelated Concepts Data)
Inventors: Hannah Cohen (Morrisville, PA, USA)
          Myles Novick (New York, NY, USA)
          Timothy Schwab (Washington, DC, USA)
Applicant: Palantir Technologies Inc. (Denver, CO, USA)
Application Number: 18/239,471
Filing Date: August 29, 2023
Patent Issuance Date: February 25, 2025
▷ Provisional Patent Application
US 63/402,599 (filed on August 31, 2022)
▷ International Patent Classification (IPC):
G06F 16/00 (Data Management Technology)
G06F 16/21 (Database Management)
▷ USA Patent Classification (CPC):
G06F 16/265 (Data Frame Extension)
G06F 16/213, G06F 16/2379 (Data Management Systems)
▷ Overview of the Invention
This patent describes a system and method for generating "Notional Data." In particular, it includes technology for associating different conditions of data and generating consistent data with each other.
⑴ Receiving and Processing Data
・Receiving "Seed Data (Initial Data)" for the Data Frame (tabular data).
Define the "object type (condition of the data)" of the seed data.
Set the "functional relationship" of that object type.
(2) Data transformation
The system generates new data based on the functional relationships.
For example, when a change in one data impacts another data, it automatically generates "new data" that reflects that impact.
(3) Examples of system application
Integration of corporate data (centrally managing data from different systems)
Creation of simulation data (creating virtual data from real data)
・Big Data analysis (extracting useful information from large amounts of data)
This technology allows the creation of datasets that more accurately Reflect the relationships between data. This plays an important role in data management in fields such as finance, healthcare, and Cybersecurity.
ーーー
Explanation of terms
▷Clustering
A method of grouping similar data. It is commonly used in Machinery learning and data analysis.
▷SEED
Information that serves as the starting point for data processing.
▷ DataFrame
A tabular data structure (composed of rows and columns).
▷ Functional Relationship
A relationship that shows how one set of data affects another.