US 12,124,925 B2
Dynamic analysis and monitoring of machine learning processes
Barum Rho, Toronto (CA); Kin Kwan Leung, Toronto (CA); Maksims Volkovs, Toronto (CA); and Tomi Johan Poutanen, Toronto (CA)
Assigned to The Toronto-Dominion Bank, Toronto (CA)
Filed by The Toronto-Dominion Bank, Toronto (CA)
Filed on Oct. 6, 2020, as Appl. No. 17/063,847.
Claims priority of provisional application 63/074,078, filed on Sep. 3, 2020.
Prior Publication US 2022/0067580 A1, Mar. 3, 2022
Int. Cl. G06N 20/00 (2019.01); G06F 9/451 (2018.01); G06N 3/08 (2023.01)
CPC G06N 20/00 (2019.01) [G06F 9/451 (2018.02); G06N 3/08 (2013.01)] 21 Claims
OG exemplary drawing
 
1. An apparatus, comprising:
a memory storing instructions;
a communications interface; and
at least one processor coupled to the memory and the communications interface, the at least one processor being configured to execute the instructions to:
receive elements of request data from a device via the communications interface, the elements of request data comprising a first identifier of a machine learning process, a second identifier of a dataset, a third identifier of a corresponding analytical period, feature data that specifies an input feature of the machine learning process, a feature range associated with the specified input feature, and a number of interpolation points associated with the specified feature range, and segmentation data that specifies a composition of a segment of the dataset;
determine a plurality of candidate feature values within the specified feature range based on the specified number of interpolation points;
obtain composition data associated with the machine learning process from the memory based on the first identifier, and obtain the dataset from the memory based on the second identifier, the composition data characterizing a composition of a feature vector associated with the machine learning process;
perform operations that access the segment of the dataset in accordance with the segmentation data, and that generate a segmented dataset that includes at least a portion of the accessed segment of the dataset;
based on the segmented dataset, generate the feature vector in accordance with the composition data, and generate a plurality of modified feature vectors based on the feature vector, each of the modified feature vectors comprising a modified feature value of the specified input feature and a corresponding one of the plurality of candidate feature values;
based on an application of the machine learning process to the feature vector and to the modified feature vectors during the corresponding analytical period, generate predictive output data associated with corresponding ones of the feature vector and the modified feature vectors and first explainability data associated with the specified input feature;
based on at least the third identifier, identify an analytical process associated with the machine learning process and the corresponding analytical period, and based on an application of the analytical process to at least a portion of the predictive output data, generate a value of one or more metrics that characterize a performance or an operation of the machine learning process during the corresponding analytical period; and
transmit the first explainability data and the one or more metric values to the device via the communications interface, the device being configured to execute an application program that presents a graphical representation of the first explainability data and the one or more metric values within a portion of a digital interface.