Patent Retrieval TRIZ Cases
- Document Entity Comparison for Patentability Assessment Accuracy
Entity recognition and hierarchical modeling improve cross-document claim matching for more accurate novelty and inventive-step assessment.
- Document Entity Comparison for Patent Dependency Analysis
Entity matching with hierarchical dependency analysis reveals cross-document correlations that improve patentability and validity assessment.
- IP Document Search with Weighted Keyword and Related-Term Ranking
Automated keyword extraction, related-term weighting, and ranking improve intellectual property document search accuracy with simple input.
- Patent Document Annotation via LLM API Mediation
An intermediary LLM API extracts special technical features from patent documents to generate annotations, tables, summaries, and claim drafts with lower integration burden.
- Biological Sequence Database Retrieval Using Attribute Association
Pre-extracted biological sequences linked to attribute data enable faster, more precise patent document retrieval for novelty and infringement searches.
- Examination Forecast Workflow Using External LLM Mediation
External LLM mediation combines search-based examination records with local processing to generate reliable refusal forecasts and argument drafts.
- Document Search Ranking Using Weighted Related Terms
Weighted keyword and related-term extraction improves IP document search accuracy while reducing manual query selection and review time.
- Analytic Product Linking Through Metadata Networks for Reproducibility
Process, session, and product nodes connect analytic results to source context, preserving provenance for reproducibility and reducing duplicate work.
- OCR Document Splitting for Omnibus File Docketing
OCR and machine learning split omnibus files into individual documents, improving docketing speed and deadline assignment.
- Interactive Patent Visualization System for Document Analysis
Dynamic graphical user interface segments complex patent documents into interactive visual components for rapid analysis.
- Document Analysis Architecture for Classification Model Training
A document analysis platform trains classification models using user input data to predict in-class or out-of-class status.
- Computer Architecture for Automated Patent Portfolio Quantification
An improved computer architecture quantifies technologies into market-tech units to resolve inefficient patent management and resource waste.
- Pre-coded Automatic Docketing for Unstructured Text Sorting
A pre-coding system categorizes incoming electronic communications into predetermined buckets using structured text identifiers.
- Machine Learning Patent Classification Model
A machine learning model learns user-defined classification standards to automate patent document sorting.
- Numeric Index for Contextual Data Retrieval
A numeric index extracts and contextualizes number-unit pairs from unstructured text.
- Generalized Data Mining Using Term Tensors for Insight Discovery
Term tensors bridge massive unstructured data and user insights by automatically identifying novel trends without predefined queries.
- Information Retrieval System Using Specification and Drawing Similarity
An information retrieval system calculates similarity degrees between application specifications and drawings to extract related products.
- Innovation Management Database for Patent Protection Planning
An improved computer system ingests documents and classifies product-tech units using machine learning algorithms.
- Adaptive Query Expansion for Cross-Lingual Document Retrieval
A search system modifies queries using analytics to enhance relevance ranking across languages.
- Weighted Section Search Profiles for Patent Document Retrieval
Profiles weight specific sections to reduce manual selection burden and improve accuracy.