Multiplex, Temporally Resolved Molecular Signal Recorder and Related Methods

EPO published patent EP4396735A1 for University of Washington covering multiplex, temporally resolved molecular signal recording systems and methods. The invention relates to biological molecular recording systems for capturing and storing molecular signals over time, classified under G06N 3/12 (AI computing) and biotechnology codes.

Simultaneous Shooting Time-Lapse Seismic Data Deblending for CO2 Monitoring

EPO published patent application EP4392812A1 for Schlumberger's AI-based seismic data deblending system designed for CO2 monitoring applications. The invention utilizes neural networks (G06N 3/02) to separate blended seismic signals in simultaneous shooting time-lapse surveys, enabling more efficient subsurface monitoring for carbon capture and storage projects.

Deep Learning for Generating Orthodontic Aligner Intermediate Stages

The European Patent Office published patent application EP4384114A1 by Solventum Intellectual Properties Company for a deep learning method generating intermediate stages for orthodontic aligners. The invention, classified under IPC A61C 7/00 and G06N 20/00, uses neural networks to predict treatment progression between initial and target tooth positions. This A1 publication covers all 31 designated EPC contracting states including EU member states.

Delivery Robot System and Method

The European Patent Office (EPO) published patent application EP4555450A1 for a delivery robot system and method using artificial intelligence. The applicant is Turkish company Delivers AI Robotik Otonom Surus Bilgi Teknolojileri A.S. with inventors Arda Agababaoglu, Ali Bolu, and Oral Yigitkus. The patent is classified under IPC G06N 3/08 (machine learning) and is designated for validation across 27+ EPO member states including Germany, France, UK, Italy, Spain, and the Netherlands.

AI/ML Channel State Information Reporting Device and Method

EPO published patent application EP4548675A1 by Shenzhen TCL New Technology Co., Ltd for a communication device and method determining channel state information using artificial intelligence and machine learning in wireless communications. The invention covers AI/ML-based techniques for channel state information reporting in telecommunications systems. Designated states include all EU member states plus other European countries.

Neural Network Patent for Industrial Equipment Lifespan Estimation

The European Patent Office published Siemens Aktiengesellschaft's patent application EP4555441A1 for a neural network-based method and apparatus for estimating remaining useful life of industrial equipment. The patent covers AI/ML techniques for predictive maintenance of machinery and industrial systems. This is an A1 publication (European patent application with search report).

QUALCOMM federated learning, clustered peer-to-peer AI systems

The European Patent Office published patent application EP4573497A1 for QUALCOMM's methods for discovery and signaling procedures in clustered and peer-to-peer federated learning AI systems. The patent covers technical procedures for distributed machine learning across telecommunications networks.

Delay Tolerant Edge Compute Protocol

The European Patent Office published patent application EP4352923A1 for Noodle Technology Inc., titled 'Delay Tolerant Edge Compute Protocol.' The patent covers communication protocols for edge computing environments with delay-tolerant networking capabilities. The filing designates all EPC contracting states including DE, FR, GB, IT, NL, and 28 other jurisdictions.

Secure Data Loading in Multiparty Computing - Royal Bank of Canada Patent

EPO published patent application EP4352681A1 by Royal Bank of Canada for a system and method for loading secure data in multiparty secure computing environments. The invention covers technical methods for secure data transfer between multiple parties in collaborative computing settings. Named inventors include Ortiz, McKay, Knöss, Baek, Khandavilli, Nabulsi, Badal-Badalian, and Simonelis.

Detecting Missing Assets Based on Work Cycle Measurements

The European Patent Office published Caterpillar Inc.'s patent EP4423687A1 on March 25, 2026, covering systems and methods for detecting missing assets based on work cycle measurements. The patent names four inventors and is classified under G06Q 10/06, G06Q 10/04, and G06Q 10/08. The designated states include all EU member states plus other European countries.