US 11,669,374 B2
Using machine-learning methods to facilitate experimental evaluation of modifications to a computational environment within a distributed system
Alexandra Savelieva, Bellevue, WA (US); Srinivas Rao Choudam, Bothell, WA (US); and Isidro Rene Hegouaburu, Kirkland, WA (US)
Filed by Microsoft Technology Licensing, LLC, Redmond, WA (US)
Filed on Aug. 10, 2022, as Appl. No. 17/884,829.
Application 17/884,829 is a continuation of application No. 16/132,212, filed on Sep. 14, 2018, granted, now 11,423,326.
Prior Publication US 2022/0383201 A1, Dec. 1, 2022
Int. Cl. G06F 9/50 (2006.01); G06N 20/00 (2019.01); G06F 17/10 (2006.01); H04L 41/0866 (2022.01); G06F 18/21 (2023.01)
CPC G06F 9/5072 (2013.01) [G06F 17/10 (2013.01); G06F 18/217 (2023.01); G06N 20/00 (2019.01); H04L 41/0866 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method for experimenting with modifications to a computational environment within a distributed system, comprising:
identifying a machine learning model trained to predict outputs for a computational environment having a first set of characteristics based on an input provided to the computational environment;
determining a modified output produced by the computational environment having a second set of characteristics, the second set of characteristics including a characteristic that has been modified from the first set of characteristics;
applying the machine learning model to the computational environment having the second set of characteristics to determine a predicted output based on the first set of characteristics; and
comparing the modified output to the predicted output to generate an output indicating an effect of a modification of the characteristic of the second set of characteristics that is different from the first set of characteristics.