US 12,137,265 B2
Machine learning for adaptive bitrate selection
Amit Paliwal, Fremont, CA (US); Andrey Marsavin, San Jose, CA (US); Govind Vaidya, San Ramon, CA (US); Wim Michiels, San Jose, CA (US); Beth Teresa Logan, Cambridge, MA (US); Zheng Han, Jersey City, NJ (US); Tapan Oza, Round Rock, TX (US); and Vijay Anand Raghavan, Newton, MA (US)
Assigned to Roku, Inc., San Jose, CA (US)
Filed by ROKU, INC., San Jose, CA (US)
Filed on Sep. 7, 2023, as Appl. No. 18/462,635.
Application 18/462,635 is a continuation of application No. 17/515,225, filed on Oct. 29, 2021.
Prior Publication US 2023/0421831 A1, Dec. 28, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. H04N 21/2662 (2011.01); G06N 5/01 (2023.01); G06N 20/00 (2019.01); G06N 20/20 (2019.01); H04N 21/24 (2011.01); H04N 21/25 (2011.01); H04N 21/44 (2011.01); H04N 21/442 (2011.01); H04N 21/845 (2011.01)
CPC H04N 21/2662 (2013.01) [G06N 5/01 (2023.01); G06N 20/00 (2019.01); G06N 20/20 (2019.01); H04N 21/2401 (2013.01); H04N 21/251 (2013.01); H04N 21/44004 (2013.01); H04N 21/44209 (2013.01); H04N 21/8456 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method for adaptive bitrate selection, comprising:
receiving, by at least one computer processor, a data streaming request;
predicting, by a speed predictive machine learning model and based on one or more streaming parameters, a current sustainable network bandwidth, wherein the speed predictive machine learning model is trained using a training data set comprising a history of the one or more streaming parameters;
predicting, by a rebuffer predictive machine learning model and based on the current sustainable network bandwidth, a buffer level of a data buffer, and a chunk duration, a candidate bitrate at which a likelihood of rebuffering of the data buffer occurs less than a threshold percentage;
selecting, based on the candidate bitrate, a download bitrate to complete the data streaming request; and
downloading streaming data at the download bitrate.