US 11,853,149 B2
Generating error event descriptions using context-specific attention
Anjali Shah, Short Hills, NJ (US); Jennifer A. Mallette, Vienna, VA (US); and Salim Roukos, Los Angeles, CA (US)
Assigned to International Business Machines Corporation, Armonk, NY (US)
Filed by International Business Machines Corporation, Armonk, NY (US)
Filed on Sep. 10, 2021, as Appl. No. 17/471,258.
Prior Publication US 2023/0084422 A1, Mar. 16, 2023
Int. Cl. G06F 11/00 (2006.01); G06F 11/07 (2006.01); G06F 40/205 (2020.01); G06F 40/284 (2020.01); G06F 11/22 (2006.01)
CPC G06F 11/0769 (2013.01) [G06F 11/2263 (2013.01); G06F 40/205 (2020.01); G06F 40/284 (2020.01)] 17 Claims
OG exemplary drawing
 
1. A computer implemented method for generating error event descriptions, the method comprising:
receiving, by one or more computer processors, a set of error messages associated with an error event;
generating, by the one or more computer processors, a tokenization of at least one line of the set of error messages;
generating, by the one or more processors, a vector from the tokenization using a linear transformation;
defining, by the one or more processors, positional information for each element of the tokenization;
providing, by the one or more computer processors, the vector and positional information of the tokenization to an encoder-decoder attention head according to a context of the tokenization;
providing, by the one or more computer processors, an output of the attention head as input to a generative model;
generating, by the one or more computer processors, a description of the error event according to the output; and
providing, by the one or more computer processors, the description to a user.