# US 2023/0274170 A1 · System for probabilistic reasoning and decision making on digital twins > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2023/0274170 A1 - Document ID: US20230274170A1 - Status: Published application - Kind: A1 - Application number: 17/681,699 - Filed: 2022-02-25 - Published: 2023-08-31 - Earliest claimed: 2022-02-25 - Assignees: Accenture Global Solutions Limited - Applicants: Accenture Global Solutions Limited - Inventors: Zaid Tashman; Matthew Kujawinski; Neda Abolhassani; Sanjoy Paul; Thien Quang Nguyen; Eric Annong Tang; Jessica Huey-Jen Yeh - Web page: https://patentgrep.com/patent/US20230274170A1 ## CPC - G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks - G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO] - G06N 3/008: based on physical entities controlled by simulated intelligence so as to replicate intelligent life forms, e.g. based on robots replicating pets or humans in their appearance or behaviour - G06N 5/02: Knowledge representation; Symbolic representation - G06N 5/022: Knowledge engineering; Knowledge acquisition - G06N 7/06: Simulation on general purpose computers - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract Aspects of the present disclosure provide systems, methods, and computer-readable storage media that support ontology driven processes to create digital twins that extend the capabilities of knowledge graphs. A dataset including an ontology and domain data corresponding to a domain associated with the ontology is obtained. A knowledge graph is constructed based on the ontology and the domain data is incorporated into the knowledge graph. The knowledge graph is exploited to derive random variables of a probabilistic graph model. The random variables may be associated with probability distributions, which may include unknown parameters. A learning process is executed to learn the unknown parameters and obtain a joint distribution of the probabilistic graph model, which may enable querying of the probabilistic graph model in a probabilistic and deterministic manner. ## Claims ### Claim 1 A method for creating digital twins, the method comprising: obtaining, by one or more processors, a dataset, wherein the dataset comprises an ontology and domain data corresponding to a domain associated with the ontology; constructing, by the one or more processors, a knowledge graph based on the ontology; incorporating, by the one or more processors, at least a portion of the domain data into the knowledge graph; and converting, by the one or more processors, the knowledge graph to a probabilistic graph model, wherein the probabilistic graph model comprises probability distributions having unknown parameters; learning, by the one or more processors, the unknown parameters based on the portion of the domain data; and updating, by the one or more processors, the probabilistic graph model based on the learning, wherein the updating replaces the unknown parameters with parameters obtained via the learning. ### Claim 2 (depends on 1) The method of claim 1, further comprising: generating a generative program based on the probabilistic graph model; and generating a set of guessed parameters corresponding to the unknown parameters, wherein the learning comprises: feeding a set of data and the set of guessed parameters to the generative program to produce simulation data; and determining whether the simulation data approximates a set of real data to within a threshold tolerance, wherein the probabilistic graph model is updated to include the parameters obtained via the learning when the simulation data approximates the set of real data to within the threshold tolerance, and wherein the parameters obtained via the learning correspond to the set of guessed parameters. ### Claim 3 (depends on 2) The method of claim 2, further comprising: modifying the set of guessed parameters in response to a determination that the simulation data does not approximate the set of real data to within the threshold tolerance; and iteratively performing the feeding, the determining, and the modifying until the simulation data approximates the set of real data to within the threshold tolerance. ### Claim 4 (depends on 2) The method of claim 2, wherein the generative program comprises deterministic statements and probabilistic statements corresponding to the probability distributions. ### Claim 5 (depends on 1) The method of claim 1, further comprising exploiting the knowledge graph to obtain random variables, each of the random variables corresponding to one of the probability distributions. ### Claim 6 (depends on 5) The method of claim 5, further comprising obtaining a joint probability distribution of the probabilistic graph model based on the probability distributions for each of the random variables. ### Claim 7 (depends on 6) The method of claim 6, further comprising: detecting a missing value in the domain data; and updating the domain data to include a value for the missing value based on a probability distribution corresponding to the missing value derived from known values associated with the missing value. ### Claim 9 (depends on 1) The method of claim 1, further comprising: generating a query associated with at least one random variable of the probabilistic graph model; and running the query against the probabilistic graph model to obtain a query result, wherein the query result comprises a distribution associated with the at least one random variable. ### Claim 10 (depends on 1) The method of claim 1, wherein the knowledge graph represents semantic relationships and the probabilistic graph model represents statistical dependencies. ### Claim 11 A system for creating digital twins, the system comprising: a memory; and one or more processors communicatively coupled to the memory, the one or more processors configured to: obtain a dataset, wherein the dataset comprises an ontology and domain data corresponding to a domain associated with the ontology; construct a knowledge graph based on the ontology; incorporate at least a portion of the domain data into the knowledge graph; convert the knowledge graph to a probabilistic graph model, wherein the probabilistic graph model comprises probability distributions having unknown parameters; learn the unknown parameters based on the portion of the domain data; and update the probabilistic graph model based on the learning, wherein the unknown parameters are replaced with parameters obtained via the learning via the update. ### Claim 12 (depends on 11) The system of claim 11, the one or more processors configured to: provide a graphical user interface providing query building functionality; receive inputs associated with random variables of the probabilistic graph model, each of the random variables corresponding to one of the probability distributions; and generate a query based on the received inputs. ### Claim 13 (depends on 12) The system of claim 12, wherein the query comprises a conditional probabilistic query regarding values of one or more of the random variables with uncertainty quantification. ### Claim 14 (depends on 11) The system of claim 11, wherein the probabilistic graph model comprises a plurality of nodes and edges connecting at least some of the plurality of nodes to one or more other nodes. ### Claim 15 (depends on 14) The system of claim 14, wherein the probability distributions correspond to random variables, and wherein each of the random variables corresponds a node of the plurality of nodes or one of the edges. ### Claim 16 (depends on 14) The system of claim 14, wherein updating the probabilistic graph model comprises embedding the probability distributions into the probabilistic graph model. ### Claim 17 A non-transitory computer-readable storage medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform operations for generating probabilistic graph models, the operations comprising: obtaining a dataset, wherein the dataset comprises an ontology and domain data corresponding to a domain associated with the ontology; constructing a knowledge graph based on the ontology; incorporating at least a portion of the domain data into the knowledge graph; and converting the knowledge graph to a probabilistic graph model, wherein the probabilistic graph model comprises probability distributions having unknown parameters; learning the unknown parameters based on the portion of the domain data; and updating the probabilistic graph model based on the learning, wherein the updating replaces the unknown parameters with parameters obtained via the learning. ### Claim 18 (depends on 17) The non-transitory computer-readable storage medium of claim 17, the operations further comprising: obtaining additional domain data; and performing additional learning based on the additional domain data. ### Claim 19 (depends on 18) The non-transitory computer-readable storage medium of claim 18, the operations further comprising modifying at least one parameter obtained via the learning based on at least one parameter obtained via the additional learning. ### Claim 20 (depends on 17) The non-transitory computer-readable storage medium of claim 17, wherein the probabilistic graph model represents a digital twin of a real world counterpart, the operations comprising: generating a query associated with a random variable of the probabilistic graph model, the random variable corresponding to one of the probability distributions; running the query against the probabilistic graph model to obtain a query result, wherein the query result comprises a distribution associated with the random variable; generating a control signal based on the query result; and transmitting the control signal to the real world counterpart.