US 11,704,540 B1
Systems and methods for responding to predicted events in time-series data using synthetic profiles created by artificial intelligence models trained on non-homogenous time series-data
Thomas Francis Gianelle, Colleyville, TX (US); Ernst Wilhelm Spannhake, II, Canal Winchester, OH (US); and Milan Shah, Plano, TX (US)
Assigned to Citigroup Technology, Inc., New York, NY (US)
Filed by Citigroup Technology, Inc., New York, NY (US)
Filed on Dec. 13, 2022, as Appl. No. 18/65,441.
Int. Cl. G06N 3/0464 (2023.01); G06N 3/045 (2023.01); G06N 3/088 (2023.01)
CPC G06N 3/0464 (2023.01) [G06N 3/045 (2023.01); G06N 3/088 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A system for responding to predicted events in computer systems using artificial intelligence models, the system comprising:
one or more processors; and
a non-transitory computer-readable medium comprising instructions that when executed by the one or more processors cause operations comprising:
generating a user profile for a user by:
receiving a first data set comprising a current state characteristic for a first system state; and
receiving a required future state characteristic for the first system state;
generating a synthetic profile for the user based on historic time-series data by:
receiving the historical time-series data;
training, using the historical time-series data, a second model using unsupervised learning, wherein the second model comprises a convolutional neural network; and
selecting, using the second model, a second data set from a plurality of available datasets based on similarities between state characteristics for the second data set and the current state characteristic and the required future state characteristic, wherein the second data set comprises second rate-of-change data over a second time period;
comparing the second rate-of-change data to a threshold rate of change to detect a rate-of-change event;
generating a normalized rate-of-change event by normalizing the rate-of-change event based on the first data set; and
using the synthetic profile to generate a recommendation for the user by:
inputting the first data set into a first model to generate first rate-of-change data over a first time period for the first system state;
generating modified first rate-of-change data based on the normalized rate-of-change event; and
generating for display, on a user interface, a recommendation based on the modified first rate-of-change data.