US 11,954,929 B2
Failure mode discovery for machine components
Rajaram Kudli, Bengaluru (IN); Satish Padmanabhan, Sunnyvale, CA (US); and Fuk Ho Pius Ng, Portland, OR (US)
Assigned to DIMAAG-AI, Inc., Fremont, CA (US)
Filed by DIMAAG-AI, Inc., Fremont, CA (US)
Filed on Mar. 17, 2023, as Appl. No. 18/186,001.
Application 18/186,001 is a continuation of application No. 17/818,373, filed on Aug. 9, 2022, granted, now 11,636,697.
Prior Publication US 2024/0054800 A1, Feb. 15, 2024
This patent is subject to a terminal disclaimer.
Int. Cl. G06V 30/19 (2022.01); G06F 16/35 (2019.01); G06V 30/182 (2022.01)
CPC G06V 30/19107 (2022.01) [G06F 16/35 (2019.01); G06V 30/1823 (2022.01)] 31 Claims
OG exemplary drawing
 
1. A method comprising:
determining via a processor a word embedding based on a plurality of text documents, the text documents including a plurality of maintenance records characterizing maintenance of mechanical components;
determining via a processor a plurality of vector representations corresponding to the plurality of maintenance records and based on a word embedding;
determining via a processor a plurality of clusters based on the plurality of vector representations, the plurality of clusters corresponding to a plurality of failure modes, each of the plurality of clusters including one or more of the text documents;
determining respective labels for the plurality of clusters based on statistical analysis of the plurality of vector representations; and
storing on a storage device a correspondence between the plurality of text documents and the plurality of failure modes.