US 12,380,015 B1
Predicting tests based on change-list descriptions
Pranay Reddy Chandram, Bangalore (IN); and Kavish Seth, Bengaluru (IN)
Assigned to SYNOPSYS, INC., Sunnyvale, CA (US)
Filed by Synopsys, Inc., Mountain View, CA (US)
Filed on Jun. 21, 2022, as Appl. No. 17/845,751.
Claims priority of provisional application 63/213,331, filed on Jun. 22, 2021.
Int. Cl. G06F 11/36 (2025.01); G06F 11/22 (2006.01); G06F 11/263 (2006.01); G06F 11/3668 (2025.01); G06N 20/00 (2019.01)
CPC G06F 11/3684 (2013.01) [G06F 11/2263 (2013.01); G06F 11/263 (2013.01); G06F 11/368 (2013.01); G06F 11/3688 (2013.01); G06F 11/3692 (2013.01); G06N 20/00 (2019.01)] 10 Claims
OG exemplary drawing
 
1. A method, comprising:
receiving training data comprising change-lists and descriptions associated with the change-lists, wherein a change-list specifies a set of changes to a design or a test case, or both, and wherein the descriptions are specified in a natural language; and
training, by a processor, a machine learning (ML) model based on the training data, wherein the ML model comprises a first neural network (NN), a second NN, and a third NN, wherein the first NN includes a first input layer and a first output layer, wherein the first input layer includes a first set of input nodes corresponding to an author name, a file name, and a set of vocabulary words or phrases, wherein the first output layer includes a first set of output nodes corresponding to a set of test categories, wherein the second NN includes a second input layer and a second output layer, wherein the second input layer includes a second set of input nodes corresponding to the author name, the file name, and a set of word or character positions in the descriptions, wherein the second output layer includes a second set of output nodes corresponding to the set of test categories, and wherein the third NN includes a third input layer and a third output layer, wherein the third input layer includes a third set of input nodes corresponding to the first set of output nodes and the second set of output nodes, and wherein the third output layer includes a third set of output nodes corresponding to the set of test categories;
receiving a first change-list and a first description associated with the first change-list for a first design, wherein the first design is an integrated circuit design;
using the trained ML model to select a first subset of test cases from a set of test cases for testing the first design based on the first change-list and the first description; and
running the first subset of test cases on a testbench which simulates the first design.