# US 2026/0086514 A1 · Deep Causal Learning for Continuous Testing, Diagnosis, and Optimization

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2026/0086514 A1
- Document ID: US20260086514A1
- Status: Published application
- Kind: A1
- Application number: 19/406,643
- Filed: 2025-12-02
- Published: 2026-03-26
- Earliest claimed: 2019-03-15
- Assignees: 3M Innovative Properties Company
- Applicants: 3M Innovative Properties Company
- Inventors: Gilles J. Benoit; Brian E. Brooks; Peter O. Olson; Tyler W. Olson
- Web page: https://patentgrep.com/patent/US20260086514A1

## CPC

- G05B 13/042: {in which a parameter or coefficient is automatically adjusted to optimise the performance}
- G05B 19/042: using digital processors (G05B19/05 takes precedence)
- B60W 40/064: Degree of grip
- B60W 40/08: related to drivers or passengers
- B60W 40/105: Speed
- G05B 13/021: {in which a variable is automatically adjusted to optimise the performance}
- G05B 13/024: {in which a parameter or coefficient is automatically adjusted to optimise the performance}
- G05B 13/0265: {the criterion being a learning criterion}
- G05B 13/041: {in which a variable is automatically adjusted to optimise the performance}
- G05B 19/4065: Monitoring tool breakage, life or condition
- G05B 19/41835: {characterised by program execution}
- G05B 23/0229: {knowledge based, e.g. expert systems; genetic algorithms}
- G05B 23/024: {Quantitative history assessment, e.g. mathematical relationships between available data; Functions therefor; Principal component analysis [PCA]; Partial least square [PLS]; Statistical classifiers, e.g. Bayesian networks, linear regression or correlation analysis; Neural networks}
- G05B 23/0248: {Causal models, e.g. fault tree; digraphs; qualitative physics}
- G06F 18/2193: {based on specific statistical tests}
- G06N 5/043: Distributed expert systems; Blackboards
- G06N 5/046: Forward inferencing; Production systems
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06N 20/00: Machine learning
- G06Q 10/06315: {Needs-based resource requirements planning or analysis}
- G06Q 10/06395: {Quality analysis or management}
- G06Q 30/0202: Market predictions or forecasting for commercial activities
- G05B 2219/36301: Optimisation of sequence of operations
- G06N 3/08: Learning methods
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06Q 10/063: Operations research, analysis or management
- G06Q 10/087: Inventory or stock management, e.g. order filling, procurement or balancing against orders
- Y02P 90/80: Management or planning
- Y02P 90/82: Energy audits or management systems therefor
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- B60W: CONJOINT CONTROL OF VEHICLE SUB-UNITS OF DIFFERENT TYPE OR DIFFERENT FUNCTION; CONTROL SYSTEMS SPECIALLY ADAPTED FOR HYBRID VEHICLES; ROAD VEHICLE DRIVE CONTROL SYSTEMS FOR PURPOSES NOT RELATED TO THE CONTROL OF A PARTICULAR SUB-UNIT
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- Y02P: CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS

## Abstract

A system and methods for multivariant learning and optimization repeatedly generate self-organized experimental units (SOEUs) based on the one or more assumptions for a randomized multivariate comparison of process decisions to be provided to users of a system. The SOEUs are injected into the system to generate quantified inferences about the process decisions. Responsive to injecting the SOEUs, at least one confidence interval is identified within the quantified inferences, and the SOEUs are iteratively modified based on the at least one confidence interval to identify at least one causal interaction of the process decisions within the system. The causal interaction can be used for testing, diagnosis, and optimization of the system performance.

## Claims

### Claim 1

A system comprising:
a memory; and
a processor in communication with the memory, the processor being configured to:
receive one or more assumptions for a randomized selection of process decisions;
based on the one or more assumptions, generate a plurality of experimental units for a treatment assignment;
inject the plurality of experimental units for the treatment assignment into the system to generate inferences about causal effects of the process decisions;
in response to injecting the plurality of experimental units into the system, identify one or more measures of precision associated with the inferences about the causal effects; and
iteratively modify the experimental units for the treatment assignment based on the one or more measures of precision to identify at least one causal interaction of the process decisions within the system.

### Claim 2 (depends on 1)

The system of claim 1, wherein to identify the one or more measures of precision associated with the inferences about the causal effects, the processor is configured to compute the one or more measures of precision using statistical testing on differences between measured effects when a treatment is activated and when the treatment is deactivated.

### Claim 3 (depends on 1)

The system of claim 1, wherein to iteratively modify the experimental units for the treatment assignment, the processor is configured to randomly select one or more characteristics of the experimental units using one of a uniform distribution, a Poisson distribution, a Gaussian distribution, or a binomial distribution.

### Claim 4 (depends on 1)

The system of claim 1, wherein each respective experimental unit of the plurality of experimental units is associated with a temporal characteristic or a spatial characteristic.

### Claim 5 (depends on 1)

The system of claim 1, wherein an experimental unit of the plurality of experimental units is associated with a temporal characteristic that represents a duration of the experimental unit being active within the system.

### Claim 6 (depends on 1)

The system of claim 1, wherein an experimental unit of the plurality of experimental units is associated with a spatial characteristic that represents a space across which the process decisions can be combined or sequenced.

### Claim 7 (depends on 1)

The system of claim 1, wherein at least one of the one or more assumptions regards a previous discovery or learning.

### Claim 8 (depends on 1)

The system of claim 1, wherein at least one of the one or more assumptions regards a previous discovery or learning obtained from observational or historical data.

### Claim 9 (depends on 1)

The system of claim 1, wherein at least one the one or more assumptions regards a previous discovery or learning obtained using an artificial intelligence or machine learning technique.

### Claim 10 (depends on 1)

The system of claim 1, wherein the processor is further configured to dynamically modify a latent duration between the plurality of experimental units to reduce a carryover effect of a prior experimental unit on a subsequent experimental unit of the plurality of experimental units.

### Claim 11

A method comprising:
receiving one or more assumptions for a randomized selection of process decisions;
generating, based on the one or more assumptions, a plurality of experimental units for a treatment assignment;
injecting the plurality of experimental units for the treatment assignment into a system of interest to generate inferences about causal effects of the process decisions;
in response to injecting the plurality of experimental units into the system of interest, identifying one or more measures of precision associated with the inferences about the causal effects; and
iteratively modifying the experimental units for the treatment assignment based on the one or more measures of precision to identify at least one causal interaction of the process decisions within the system of interest.

### Claim 12 (depends on 11)

The method of claim 11, wherein identifying the one or more measures of precision comprises performing statistical testing on corresponding d-scores, and wherein each respective d-score represents a difference in a measured effect between a level of an independent variable of the treatment assignment being active and the level of the independent variable of the treatment assignment not being active.

### Claim 13 (depends on 11)

The method of claim 11, wherein at least one of the one or more assumptions regards a previous discovery or a previous learning.

### Claim 14 (depends on 11)

The method of claim 11, wherein at least one of the one or more assumptions regards a previous discovery or learning obtained from observational or historical data.

### Claim 15 (depends on 11)

The method of claim 11, wherein at least one the one or more assumptions regards a previous discovery or learning obtained using an artificial intelligence or machine learning technique.

### Claim 16 (depends on 11)

The method of claim 11, further comprising grouping the plurality of experimental units into two or more clusters based on quantified inferences of variance in the causal effects of the process decisions.

### Claim 17 (depends on 16)

The method of claim 16, further comprising, for each respective cluster of the two or more clusters, implementing blocking to reduce a variability between the respective experimental units of each respective cluster.

### Claim 18 (depends on 17)

The method of claim 17, wherein implementing the blocking comprises assigning a presence of a control element and an absence of a control element to respective ones of at least two experimental units of each respective cluster, the method further comprising selecting the at least two experimental units by applying propensity matching to the respective experimental units of each respective cluster.

### Claim 19 (depends on 16)

The method of claim 16, wherein grouping the plurality of experimental units into the two or more clusters comprises forming the two or more clusters based on at least one inter-cluster difference between the two or more clusters as identified using one or more of a factorial ANOVA technique, an independence testing technique, a conditional inference tree-based technique, or an unsupervised classification technique.

### Claim 20

An apparatus comprising:
means for receiving one or more assumptions for a randomized selection of process decisions;
means for generating a plurality of experimental units for a treatment assignment based on the one or more assumptions;
means for injecting the plurality of experimental units for the treatment assignment into a system of interest to generate inferences about causal effects of the process decisions;
means for identifying one or more measures of precision associated with the inferences about the causal effects in response to the injection of the plurality of experimental units into the system of interest; and
means for iteratively modifying the experimental units for the treatment assignment based on the one or more measures of precision to identify at least one causal interaction of the process decisions within the system of interest.

### Claim 21

A non-transitory computer-readable medium encoded with instructions that, when executed, cause a processor of a system to:
receive one or more assumptions for a randomized selection of process decisions;
generate, based on the one or more assumptions, a plurality of experimental units for a treatment assignment;
inject the plurality of experimental units for the treatment assignment into a system of interest to generate inferences about causal effects of the process decisions;
identify one or more measures of precision associated with the inferences about the causal effects in response to injecting the plurality of experimental units into the system of interest; and
iteratively modify the experimental units for the treatment assignment based on the one or more measures of precision to identify at least one causal interaction of the process decisions within the system.

### Claim 22 (depends on 21)

The non-transitory computer-readable medium of claim 21, wherein at least one of the one or more assumptions regards a previous discovery or learning.

### Claim 23 (depends on 22)

The non-transitory computer-readable medium of claim 22, wherein at least one of the one or more assumptions regards a previous discovery or learning obtained from observational or historical data.

### Claim 24 (depends on 21)

The non-transitory computer-readable medium of claim 21, wherein at least one the one or more assumptions regards a previous discovery or learning obtained using an artificial intelligence or machine learning technique.

### Claim 25 (depends on 21)

The non-transitory computer-readable medium of claim 21, further encoded with instructions that, when executed, cause the processor to dynamically modify a latent duration between the plurality of experimental units to reduce a carryover effect of a prior experimental unit on a subsequent experimental unit of the plurality of experimental units.

### Claim 26 (depends on 21)

The non-transitory computer-readable medium of claim 21, wherein the instructions that cause the processor to identify the one or more measures of precision comprise instructions that, when executed, cause the processor to perform statistical testing on corresponding d-scores, and wherein each respective d-score represents a difference in a measured effect between: one level of two or more levels associated with a discretized continuous independent variable of the treatment assignment, and remaining levels of the two or more levels.

### Claim 27 (depends on 26)

The non-transitory computer-readable medium of claim 26, wherein the two or more levels include an active level of the discretized continuous independent variable, an inactive level of the discretized continuous independent variable, and at least one additional level of the discretized continuous independent variable.

### Claim 28 (depends on 21)

The non-transitory computer-readable medium of claim 21, further encoded with instructions that, when executed, cause the processor to determine each respective causal effect of the process decisions by measuring an effect of the respective process decision on a dependent variable that characterizes a response of the system of interest to the respective process decision.

### Claim 29 (depends on 28)

The non-transitory computer-readable medium of claim 28, wherein the dependent variable represents at least one of a direct measure of utility, an input to a complex utility function, or a state of the system of interest in response to the respective process decision.

### Claim 30 (depends on 28)

The non-transitory computer-readable medium of claim 28, wherein the dependent variable represents an intermediate goal associated with a set of the process decisions, wherein the set of the process decisions are arranged sequentially.
