US 11,989,517 B2
Conversational automated machine learning
Meenakshi Sundaram P, Tirunelveli (IN); and Gokulraj Ramdass, Walldorf (DE)
Assigned to SAP SE, Walldorf (DE)
Filed by SAP SE, Walldorf (DE)
Filed on Sep. 13, 2021, as Appl. No. 17/473,891.
Prior Publication US 2023/0078800 A1, Mar. 16, 2023
Int. Cl. G06F 40/30 (2020.01); G06F 9/48 (2006.01); G06F 18/21 (2023.01); G06F 40/295 (2020.01); G06F 40/40 (2020.01); G06N 20/00 (2019.01)
CPC G06F 40/295 (2020.01) [G06F 9/4881 (2013.01); G06F 18/2178 (2023.01); G06F 40/40 (2020.01); G06N 20/00 (2019.01)] 20 Claims
OG exemplary drawing
 
15. A computer-implemented method, comprising:
identifying a dataset based on a user selection;
determining a machine learning problem type and an analysis type based on user input provided to a conversational interface;
determining one or more machine learning algorithms based on the dataset, the machine learning problem type, and the analysis type;
generating a routing slip for each of the one or more machine learning algorithms based on the dataset, the machine learning problem type, and the analysis type, each routing slip specifying a sequence of processing steps based on the dataset and a particular machine learning algorithm, the sequence of processing steps including an exploratory data analysis step, a machine learning parameter selection step and a machine learning model training step;
performing the sequence of processing steps specified in the routing slip for each of the one or more machine learning algorithms to generate one or more machine learning models, the machine learning model training step generating a machine learning model using parameters selected in the machine learning parameter selection step;
receiving, before completion of the sequence of processing steps and after review of the exploratory data analysis step by a user, a request to return a particular step in the sequence of processing steps specified in the routing slip, the request comprising one or more new parameters received at the conversational interface;
in response to the request, returning to the particular step specified in the routing slip;
processing the particular step according to the one or more new parameters;
continuing performing the sequence of processing steps after the particular step according to the sequence of steps specified in the routing slip;
determining an accuracy score for each of the one or more machine learning models; and
presenting the accuracy score for each of the one or more machine learning models to the user via the conversational interface.