Nvidia patents search system that favors answers from subject experts
Nvidia describes a system that improves how AI assistants find information by checking who wrote the source material. It ranks documents higher if the author is identified as an expert on the specific topic, then filters out less relevant results before summarizing them for a final answer.
In the filing’s own words
In various examples, an enhanced retrieval-augmented generation (RAG) system may provide relevant responses to user queries by prioritizing contextually relevant information based on author-related metadata. For instance, the systems and methods of the present disclosure may enrich documents—or portions (e.g., “chunks”) thereof—with metadata indicating an author's role, expertise, or association with a customer or topic. Upon receiving a query, the RAG system may retrieve content using keyword and/or semantic search, and prioritize results based on expert relevance. In some examples, threshold-based filtering or expert relationship graphs may be used to improve retrieval quality and reduce processing costs. Retrieved documents may, in some instances, be broken into smaller parts and summarized in parallel, with intermediate summaries recursively condensed to fit within a language model's context window. A final answer may then be generated using the user query and the refined summary.
What exactly they claim (claim 1)
1. A method comprising: obtaining, from one or more databases at least partially responsive to receiving input data representing a query related to a subject, a set of documents determined to be relevant to the query based at least on a semantic similarity between a vector representation of the query and vector representations of the documents; ranking the documents based at least on author information stored in association with the documents such that documents authored by individuals identified as experts with respect to the subject of the query are ranked higher than documents authored by non-experts; updating the set of documents as ranked by removing at least one document having an alignment score that is less than a threshold, the alignment score indicating a degree of association between an author of the document and the subject of the query; and generating, using one or more language models, a response to the query based at least on a summary of the set of documents as updated.
Facts
- Company
- Nvidia (NVIDIA Corporation)
- Official title
- EXPERT-AWARE RETRIEVAL AUGMENTED GENERATION
- Publication number
- US20260277980A1
- Application number
- 19405244
- Filed
- Dec 1, 2025
- Published
- Sep 17, 2026
- Inventors
- 11
- Claims
- 20
Classified under
G06F 16/345 Physics & computingG06F 16/3347 Physics & computingG06F 16/35 Physics & computingG06F 16/38 Physics & computingMore from Nvidia in AI & software
- Nvidia patents AI coach that analyzes your gameplay to suggest strategies Oct 13, 2026
- Nvidia patents AI system that writes and voices character conversations Sep 10, 2026
- Nvidia patents system that uses AI to find bugs in game videos Sep 17, 2026
- Nvidia patents AI method to make digital characters move realistically Sep 8, 2026
- Nvidia patents AI system that builds image generators from text prompts Sep 15, 2026
Summaries are simplified from the filing’s own abstract and first claim. A patent filing is not a product announcement.
Patent data and drawings: U.S. Patent and Trademark Office (USPTO). Nebula is not affiliated with or endorsed by the USPTO.