Identity Graph Fraud Detection Systems and Methods - Stripe Patent EP4172908A1

The European Patent Office published Stripe LLC's patent application EP4172908A1 for identity graph-based fraud detection systems and methods. The invention covers AI/machine learning techniques for analyzing identity relationships to detect fraudulent activity across payment networks. The patent is designated for protection in all 35 European states including Germany, France, UK, Italy, Spain, and the Netherlands.

A Machine Learning Framework and Method for Using the Same

The European Patent Office published patent application EP4172902A1 for Stripe, LLC covering a machine learning framework and method for payment and financial applications. The patent names six inventors and designates 32 European states. This publication makes the technical disclosure publicly available and establishes priority dates for the claimed ML innovations.

ZTE Federated Learning Participant Selection Patent, G06N 20/00

The European Patent Office granted patent EP4250188A1 to ZTE Corporation covering federated learning participant selection methods and apparatus. The patent protects ZTE's technology for intelligently selecting participants in federated learning systems based on model performance metrics and communication efficiency. This patent is effective across all designated EPO member states including Germany, France, the UK, Italy, Spain, and 25 other European countries.

Object Recognition Neural Network for Amodal Center Prediction

The European Patent Office published patent EP4172862A1 for Magic Leap, Inc., covering a neural network method and system for predicting amodal center points to determine complete object shapes from partial visual observations in object recognition applications.

Panomic genomic prevalence score, Caris MPI, Inc., EPO A1, 1st Apr

Panomic genomic prevalence score, Caris MPI, Inc., EPO A1, 1st Apr

Method for Assigning a Vertigo Patient to a Medical Specialty

The European Patent Office published patent EP4233069A1 for Vertify GmbH, covering a method for assigning vertigo patients to appropriate medical specialties using health informatics algorithms. The patent, classified under IPC G16H 50/20 and G16H 30/40, is valid across 31 designated European states including all EU member states and EPC contracting states.

Context-Adaptive Behavior Determination System for Health

The EPO published patent application EP4244866A1 by Hochschule Düsseldorf (Germany) covering methods and systems for assisting end users in determining context-adaptive behavior of a system. The invention relates to health informatics technology under IPC classification G16H 40/67.

System and Method for Classifying and Using Chronotypes

The European Patent Office granted ResMed Patent EP4243692A1 covering a system and method for classifying chronotypes using biometric data analysis. The patent application was published in April 2026 and names Matthew Shaw, David Zeid Aridi, Nicholas James Mealey, Jessica Xu, and Varuni Lakshana Vithanage Fernando as inventors. The patent covers designated European states including Germany, France, the United Kingdom, Italy, Spain, and other EU member states.

Classifying Images of Dose-Response Graphs - EP4252248A1

The European Patent Office published Sanofi patent EP4252248A1 for classifying images of dose-response graphs using automated image analysis in health informatics applications (G16H 20/10). The patent covers systems and methods for analyzing pharmaceutical dose-response data through image classification. Valid in designated European states including France, Germany, United Kingdom, Italy, and Spain.

ResMed Patent EP4248458A1 - Comfort Monitoring System

The EPO granted ResMed Sensor Technologies Limited a new patent (EP4248458A1) for a system and method of monitoring an individual's comfort level, classified under Health Informatics (G16H). The patent names Roxana Tiron, Michael Scannell, Kieran Conway, and Redmond Shouldice as inventors, and is designated across all EU member states.