This invention describes an automated method to build a structured vocabulary, or 'ontology,' for a Virtual Assistant that works in multiple languages. It analyzes words from the Virtual Assistant's web pages, identifies their grammatical roles and base forms, and then organizes this linguistic data into a network. The claims specify that this process involves extracting information from various document types using specialized software.
Why it matters: Filed before the widespread adoption of large language models. Modern NLP tools and LLMs can now automate and improve the multi-language linguistic analysis and ontology generation steps, which were more complex in 2015.
AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.
Reinvent this with AI