US 12,354,016 B2
Parallel inference processing by decision tree leaf nodes
Charles E. Hackett, Greenlawn, NY (US)
Assigned to INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed by INTERNATIONAL BUSINESS MACHINES CORPORATION, Armonk, NY (US)
Filed on Oct. 25, 2021, as Appl. No. 17/509,180.
Prior Publication US 2023/0128776 A1, Apr. 27, 2023
Int. Cl. G06N 5/01 (2023.01); G06N 20/00 (2019.01); G06N 20/20 (2019.01)
CPC G06N 5/01 (2023.01) [G06N 20/00 (2019.01); G06N 20/20 (2019.01)] 20 Claims
OG exemplary drawing
 
1. A system for facilitating processing within a computing environment, the system comprising:
a memory; and
a processing circuit in communication with the memory, wherein the system is configured to perform a method, the method comprising:
providing a decision tree inference accelerator to perform inference processing, the decision tree inference accelerator being configured to perform operations comprising:
generating obtaining, by the decision tree inference accelerator, a decision node result vector based on processing of an input data vector through decision nodes of a decision tree; and
parallel processing, by the decision tree inference accelerator, the decision node result vector, based on respective, predetermined path vector through the decision tree for each leaf node of a plurality of leaf nodes, the parallel processing of the decision node result vector using respective masks for each leaf node determined from the respective, predetermined path vector through the decision tree for that leaf node, wherein using the respective masks only one leaf node is selected for the decision tree for any given input data vector during the inference processing, thereby allowing the decision node result vector to be processed by the decision tree inference accelerator in parallel through the plurality of leaf nodes and determining therefrom which leaf node of the plurality of leaf nodes is selected.