US 12,288,410 B2
Activity classification using unsupervised machine learning
Anupam Dewan, Seattle, WA (US); Mengyuan Liu, Bothell, WA (US); and Weikun Hu, Seattle, WA (US)
Assigned to SAP SE, Walldorf (DE)
Filed by SAP SE, Walldorf (DE)
Filed on Oct. 4, 2022, as Appl. No. 17/959,620.
Prior Publication US 2024/0112481 A1, Apr. 4, 2024
Int. Cl. G06V 30/14 (2022.01); G06N 20/20 (2019.01); G06V 30/19 (2022.01)
CPC G06V 30/19107 (2022.01) [G06N 20/20 (2019.01); G06V 30/14 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A system comprising:
a memory storing processor-executable program code; and
a processing unit to execute the processor-executable program code to cause the system to:
acquire a first image of a first activity record;
determine first text based on the first image;
generate a first embedding based on the first text;
generate a second embedding based on the first embedding using a first model, where a number of dimensions of the second embedding is less than a number of dimensions of the first embedding;
determine a first cluster based on the second embedding using a second trained model, the second trained model trained using unsupervised learning; and
determine a first activity type associated with the first activity record based on the first cluster, the second embedding and historical activity data associating the first cluster with a plurality of activity types and each of the plurality of activity types with a respective embedding metric.