This system reads text messages exchanged between various individuals or groups. It uses advanced computer models to understand the meaning of the text, creating detailed notes (annotations) about language patterns. The system then uses these annotated messages to identify relationships between the communicating parties and to build a knowledge base, learning and improving by correcting its own annotation errors and refining its underlying language models. The first claim specifically focuses on using this process to determine relationships between entities from message data.
Why it matters: Filed just as LLMs were becoming widely capable. The system's reliance on statistical language models for semantic annotation and self-correction is now significantly more powerful and accessible due to rapid advancements in LLM technology and fine-tuning since 2023.
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