Samsung patents method to help voice assistants understand follow-up questions
Samsung describes a system that helps voice assistants understand short follow-up commands by linking them to the previous conversation. It uses a specific type of neural network to compare the new input against known categories, factoring in the context of what was said before to determine the correct intent.
In the filing’s own words
A method includes receiving an input utterance that is a continuation of a previous utterance. The method also includes, using a trained Siamese network, determining input utterance embeddings representing tokens from the input utterance, pooling the input utterance embeddings with a context token embedding representing a class associated with the previous utterance to generate a representative input utterance embedding, and determining a representative embedding associated with each of multiple possible classes. Each possible class is associated with first and second threshold boundaries. The method further includes, using the trained Siamese network, determining a similarity score for each possible class based on a distance between the representative input utterance embedding and a selected threshold boundary of the representative embedding for that possible class and identifying a class for the input utterance based on the determined similarity scores. In addition, the method includes performing an action corresponding to the identified class.
What exactly they claim (claim 1)
1. A method comprising: receiving an input utterance that is a continuation of a previous utterance; using a trained Siamese network: determining input utterance embeddings representing tokens from the input utterance; pooling the input utterance embeddings with a context token embedding representing a class associated with the previous utterance to generate a representative input utterance embedding, wherein the pooling comprises assigning attention weights to the input utterance embeddings based on the context token embedding; determining a representative embedding associated with each of multiple possible classes, wherein each possible class is associated with (i) a first threshold boundary that encompasses embeddings of utterances that specify that possible class and (ii) a second threshold boundary that encompasses embeddings of continuation utterances with context, wherein determining the representative embedding includes selecting the representative embedding associated with each of the multiple possible classes as a class target, and wherein the class target represents a centroid embedding for an associated class of the multiple possible classes; determining a similarity score for each possible class based on a distance between the representative input utterance embedding and a selected threshold boundary of the representative embedding for that possible class, the selected threshold boundary comprising the first threshold boundary or the second threshold boundary; and identifying a class for the input utterance based on the determined similarity scores; and performing an action corresponding to the identified class.
Facts
- Company
- Samsung (Samsung Electronics Co., Ltd.)
- Official title
- System and method for context insertion for contrastive Siamese network training
- Patent number
- US12801849B2
- Application number
- 18315931
- Filed
- May 11, 2023
- Granted
- Nov 17, 2026
- Inventors
- 5
- Claims
- 21
Classified under
G06F 40/284 Physics & computingG10L 15/063 Physics & computingG10L 15/183 Physics & computingG10L 2015/0636 Physics & computingMore from Samsung in AI & software
- Samsung patents phone feature that edits photos using live camera view Oct 27, 2026
- Samsung patents tool to erase objects from photos using AI Sep 24, 2026
- Samsung patents drawing objects into photos with AI Oct 15, 2026
- Samsung patents phone reminders that use AI to predict what you forgot Oct 29, 2026
- Samsung patents AI agent that listens when you look at it Oct 27, 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.