System for automated similarity analysis of technical intellectual property documents using artificial intelligence

Abstract

System for analyzing technical intellectual property documents using AI, comprising: a) input of a technical description, b) generation of a semantic representation, c) comparison with existing intellectual property rights, d) evaluation and output of the similarity.

Description

1. Technical field The invention relates to the automated search and analysis of technical intellectual property rights (in particular patents and utility models) using artificial intelligence (AI), in particular natural language processing (NLP) and semantic vectorization, to identify similarities, differences and potential conflicts in technical documents. 2. State of the art Current search systems such as DEPATISnet, Espacenet, or Google Patents are primarily based on keyword searches, classification codes (e.g., IPC / CPC), or structural metadata filtering. Semantic analysis based on the technical content is not performed. This leads to incomplete or irrelevant results, especially with non-standardized terminology or synonyms. 3. Object of the invention The aim is to create a system that allows users to automatically compare their technical descriptions with existing intellectual property rights in a content-based manner. The goal is to increase the relevance and accuracy of searches, in particular through: • Identification of semantic similarities and differences, • Classification according to technical significance, • Prioritization of potential conflicts. 4. Solution to the task The task is solved by a computer-aided system that: 1. Receives a technical text (e.g., a description of an invention), 2. Performs a semantic analysis of the text, including the identification of technical features, terms, and functions, 3. Creates vector representations (embeddings) of the content, 4. Compares these with an existing patent or utility model corpus, 5. Scores and ranks the results according to content similarity, technical proximity, and conflict potential, 6. Presents the results in a structured format (e.g., table or summary report). 5. Example of implementation A web or software tool accepts free text input from the user, e.g., "System for energy-efficient wireless temperature measurement." The system normalizes the text, analyzes its structure, and generates semantic embedding. A patent database-supported similarity analysis is then performed. The most similar patents are displayed, along with their similarity metrics, commonalities, and differences. Optionally, the results can be exported as a PDF. 6. Advantages of the invention • More precise, well-founded search results • Consideration of terms that are not exactly identical (synonyms, differing formulations) • Time savings • Reduction of risks in intellectual property applications • Suitable for individual inventors, companies and patent attorneys

Claims

System for analyzing technical intellectual property documents using AI, comprising: a) input of a technical description, b) generation of a semantic representation, c) comparison with existing intellectual property rights, d) evaluation and output of the similarity. System according to claim 1, wherein a text embedding is used as the semantic representation. System according to claim 1 or 2, wherein the comparison documents are reduced in advance by multi-stage filtering according to combinations of terms. System according to one of the preceding claims, wherein the results are presented in a visually interpretable structure.

Images