Determining conversation analysis indicators for a multiparty conversation

This invention analyzes conversations involving at least two people. It takes in video, sound, and text data for each person's spoken parts. It then processes this data for each person to create individual outputs, which are combined for a specific time window. These combined features are then fed into a machine learning system to generate indicators about the conversation.

Why it matters: While the core technology existed in 2024, the rapid advancements in multimodal foundation models and their accessibility since then could significantly simplify the development and improve the accuracy of the conversation analysis indicators.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedMultimodal machine learning, data pipelines

3 AI reinvention directions for this patent

See all three

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

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

This one

+10