Amazon patents system that resumes old chats without you asking
This filing describes a method for voice assistants to automatically pick up previous conversations. It compares your new spoken request to past chats to see if they are related, then feeds that old context to the language model to help it understand your current question. If your new input matches multiple old chats, the system asks you to choose which one to continue.
In the filing’s own words
Techniques for enabling a user to resume a user-system dialog, without the user explicitly requesting resumption of the dialog, are described. The system may determine the semantic similarity of a received user input to user-system dialogs involving the user. If the system determines the user input is not semantically similar to any stored dialogs, the system may process the user input without using a dialog as context. Conversely, if the system determines the user input is semantically similar to a stored dialog, the system may use the semantically similar dialog as context when processing the user input. If the system determines the user input is semantically similar to more than one stored dialog, the system may request the user select one of the semantically similar dialogs to be used as context when processing the user input.
What exactly they claim (claim 1)
1. A computer-implemented method comprising: receiving, from a first user device, input audio data including a spoken natural language input; generating text data corresponding to a transcription of the spoken natural language input; generating, using the text data, first embedding data corresponding to a semantic representation of the spoken natural language input; determining a user identifier associated with the input audio data; determining second embedding data associated with the user identifier, wherein the second embedding data corresponds to a semantic representation of a summary of a user-system dialog that was performed using a second user device; determining a value representing a semantic similarity between the first embedding data and the second embedding data; determining the value satisfies a threshold value; based on the value satisfying the threshold value, generating prompt data including: a first portion corresponding to the text data; and a second portion corresponding to the user-system dialog; processing, using a language model, the prompt data to determine a response to the spoken natural language input, wherein the language model uses the user-system dialog as context for processing the text data; and causing presentation of the response.
Facts
- Company
- Amazon (Amazon Technologies, Inc.)
- Official title
- DIALOG CONTEXT FOR GENERATIVE MODEL PROCESSING
- Publication number
- US20260301737A1
- Application number
- 19093996
- Filed
- Mar 28, 2025
- Published
- Oct 1, 2026
- Inventors
- 5
- Claims
- 20
Classified under
G10L 15/183 Physics & computingG10L 15/22 Physics & computingMore from Amazon in AI & software
- Amazon patents system that cites sources for AI answers Sep 29, 2026
- Amazon patents tech to swap words in movies using actor's voice Sep 8, 2026
- Amazon patents ML tool to add graphics to live sports from moving cameras Sep 15, 2026
- Amazon patents shopping assistant that remembers your search filters Oct 6, 2026
- Amazon patents system where AI agents split tasks and talk to each other Oct 20, 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.