Rehovot, Israel

Ehud Karpas

This inventor holds 1 USPTO granted patent. Top assignee: Ai21 Labs. Active years: 2026.

IDiyas Innovation Intelligence. (2026). Inventor Profile: Ehud Karpas. Retrieved from https://idiyas.com/inventor/ehud-karpas

Data source: USPTO (United States Patent and Trademark Office) public patent records. Weekly synchronization. How IDiyas builds this profile

Calculating rank

% Patents Active = 100.0

Average Co-Inventor Count = 18.0

ph-index = 1


Company Filing History:


Years Active: 2026

Loading Chart...
1 patent (USPTO):Explore Patents

Title: Ehud Karpas: Inventor in Modular Reasoning and Language Systems

Introduction

Ehud Karpas is an inventor based in Rehovot, Israel, whose work includes a patent for a modular approach to reasoning, knowledge retrieval, and natural-language systems. He is associated with AI21 Labs, a company focused on artificial intelligence and language technologies.

Latest Patents

Karpas is listed as an inventor on 1 patent, titled “Modular Reasoning, Knowledge, and Language Systems.” The patent describes a computer-implemented method that uses a trained language model to break a natural-language question into multiple information requests. The system routes those requests to information resources, receives corresponding responses, and uses them to generate a natural-language answer for the user.

Career Highlights

Karpas’s patent reflects an approach that combines language models with information retrieval and modular reasoning. By dividing a question into smaller requests and drawing on relevant information resources, the disclosed system aims to produce responses grounded in multiple information inputs. Karpas is associated with AI21 Labs.

Collaborations

Karpas’s listed coworkers include Gal Shachaf and Erez Schwartz. Their shared association with AI21 Labs places his work within a broader team working in artificial intelligence and language technology.

Conclusion

Ehud Karpas’s patent on modular reasoning, knowledge, and language systems highlights an approach to answering natural-language questions by combining language models with information resources.

Profile summary based on public USPTO records.
Data Sources: USPTO Patent Grant XML, Patent Center, EPO & CIPO • Normalized by IDiyas Innovation Graph. Methodology & provenance architecturePlease report any incorrect information to [email protected]
Loading…