IP Library Granted Patent US 12705029
Granted Patent B2
US 12705029 · App. 17/954,481 · Granted Aug 11, 2026

Methods and systems of sequential latent inference of complex knowledge

Inventors: Phillip Williams (Ottawa, CA); Sebastien Ouellet (Ottawa, CA); Nathaniel Stanley (Ottawa, CA); Chantal Bisson-Krol (Ottawa, CA)
Assignee: Kinaxis Inc.
G06F8/315G06F8/75G06F11/3612G06F11/3636G06F11/3644G06Q10/08
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12705029
App. No.
17/954,481
Granted
Aug 11, 2026
Kind
B2
Abstract

Methods and systems that allow for supply chain logic to be instrumented in such a way that a supply planner can see the major factors driving KPIs, as well as drill down to the lower level to see the impact of each item at the smallest possible level.

Claims (38)

1 . A computer-implemented method comprising:

instrumenting, by an auto-differentiation engine, a codebase;

capturing during execution of the codebase, by the auto-differentiation engine, interactions of a plurality of variables in the codebase;

executing, by a processor, the codebase, thereby generating a metric;

extracting, by the processor, low-level data from the codebase, wherein the low-level data comprises dense data of the interactions of the plurality of variables; and

determining, by the processor, an impact of each of the plurality of variables on the metric, wherein the impact for each of the plurality of variables indicates how iterative changes to that one of the plurality of variables affects the metric so that the processor can compute the incremental differences in the iterative changes at each iteration and the effect of the iterative changes on a final state of the metric.

2 . The computer-implemented method of claim 1 , wherein the codebase is a C++ codebase.

3 . The computer-implemented method of claim 1 , wherein the codebase is related to supply chain management, one or more extensions are made to the auto-differentiation engine, and hooks are added to one or more points in the codebase.

4 . The computer-implemented method of claim 1 , wherein the low-level data comprises data structures, lists of executed mathematical operations, and at least one of derivatives-related data, calculus-related data, and optimization-related data.

5 . The computer-implemented method of claim 1 , wherein the low-level data comprises a directed acyclic graph.

6 . The computer-implemented method of claim 5 , wherein path-integrated gradients are used to compute the impact based on the directed acyclic graph.

7 . The computer-implemented method of claim 1 , wherein the interactions of the plurality variables in the codebase include how the plurality of the variables interact with each other, and the method further includes:

determining by the processor, how one or more of the plurality of variables impacts another one or more of the plurality of variables.

8 . The method of claim 7 , wherein the method further comprises: automatically determining an entire dependency of all the plurality of variables in a workflow based on the interactions of the plurality of variables captured during the instrumentation of the codebase.

9 . A system comprising:

a processor; and

a memory storing instructions that, when executed by the processor, configure the system to:

instrument, by an auto-differentiation engine, a codebase;

capture during execution of the codebase, by the auto-differentiation engine, interactions a plurality of variables in the codebase;

execute, by the processor, the codebase, thereby generating a metric;

extract, by the processor, low-level data from the codebase, wherein the low-level data comprises dense data of the interactions of the plurality of variables; and

determine, by the processor, an impact of each of the plurality variables on the metric, wherein the impact for each of the one or more variables indicates how iterative changes to that one of the plurality of variables affects the metric so that the processor can compute the incremental differences in the iterative changes at each iteration and the effect of the iterative changes on a final state of the metric.

10 . The system of claim 9 , wherein the codebase is a C++ codebase.

11 . The system of claim 9 , wherein the low-level data comprises data structures, lists of executed mathematical operations, and at least one of derivatives-related data, calculus-related data, and optimization-related data.

12 . The system of claim 9 , wherein the codebase is related to supply chain management, one or more extensions are made to the auto-differentiation engine, and hooks are added to one or more points in the codebase.

13 . The system of claim 9 , wherein the low-level data comprises a directed acyclic graph.

14 . The system of claim 13 , wherein path-integrated gradients are used to compute the impact based on the directed acyclic graph.

15 . A non-transitory computer-readable storage medium, the computer-readable storage medium including instructions that when executed by a computer, cause the computer to:

instrument, by an auto-differentiation engine, a codebase;

capture during execution of the codebase, by the auto-differentiation engine, interactions of a plurality of variables in the codebase;

execute, by a processor, the codebase, thereby generating a metric;

extract, by the processor, low-level data from the codebase, wherein the low-level data comprises dense data of the interactions of the plurality of variables;

determine, by the processor, an impact of each of the plurality variables on the metric, wherein the impact for each of the plurality of variables indicates how iterative changes to that one of the plurality of variables affects the metric so that the processor can compute the incremental differences in the iterative changes at each iteration and the effect of the iterative changes on a final state of the metric.

16 . The computer-readable storage medium of claim 15 , wherein the codebase is a C++ codebase.

17 . The computer-readable storage medium of claim 15 , wherein the low-level data comprises data structures, lists of executed mathematical operations, and at least one of derivatives-related data, calculus-related data, and optimization-related data.

18 . The computer-readable storage medium of claim 15 , wherein the codebase is related to supply chain management, one or more extensions are made to the auto-differentiation engine, and hooks are added to one or more points in the codebase.

19 . The computer-readable storage medium of claim 15 , wherein the low-level data comprises a directed acyclic graph.

20 . The computer-readable storage medium of claim 19 , wherein path-integrated gradients are used to compute the impact based on the directed acyclic graph.