Turn abandoned patents
into your own filing
into your own filing
This invention describes a search engine that understands the meaning of your search queries and the text it searches, rather than just matching keywords. It works by first gathering information from the internet, then analyzing it by language and enriching it with a pre-built, off-line knowledge base of meanings. This enriched data is then organized based on questions people might ask. When you type a query, the system analyzes its meaning, specifically looking for a question structure, and then scores the collected information based on how well it matches those anticipated questions, or even the website's address and title. It then shows you related questions as clickable links to continue your search.
Why it matters: Filed before Large Language Models (LLMs) existed, which now excel at understanding meaning and generating related queries. The complex 'off-line ontological semantics' and 'anticipated queries' could be implicitly handled by modern AI, making the core concept far more achievable and powerful.
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
