US 12,445,687 B1
Supplemental content recommender
Yan Fu, Bellevue, WA (US); Sunny Chiu Webster, Woodinville, WA (US); Kellen Gillespie, Bellevue, WA (US); Zeynabalsadat Raeesy, Seattle, WA (US); Qian You, Bellevue, WA (US); and Yao Zhang, Sunnyvale, CA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Mar. 19, 2024, as Appl. No. 18/609,541.
Int. Cl. H04N 21/478 (2011.01); G10L 15/22 (2006.01); H04W 8/18 (2009.01)
CPC H04N 21/478 (2013.01) [G10L 15/22 (2013.01); H04W 8/18 (2013.01)] 20 Claims
OG exemplary drawing
 
13. A computing system comprising:
at least one processor; and
at least one memory comprising instructions that, when executed by the at least one processor, cause the computing system to:
receive usage data associated with a user identifier;
receive supplemental data including supplemental content and indicating how the supplemental content is to be presented;
determine first embedding data representing the usage data;
determine second embedding data representing the supplemental data;
receive input data associated with the user identifier, the input data representing a user input;
receive response data representing a system response to the user input;
receive supplemental content data corresponding to the supplemental content;
determine, using a language model, third embedding data representing the input data;
determine, using the language model, fourth embedding data representing the response data;
determine, using the language model, fifth embedding data representing the supplemental content data;
determine model input data using the first embedding data, the second embedding data, the third embedding data, the fourth embedding data, and the fifth embedding data;
process, using a second model, the model input data to determine the supplemental content is to be presented; and
present the supplemental content.