# US 2018/0322395 A1 · Training and Control System for Evolving Solutions to Data-Intensive Problems Using Nested Experience-Layered Individual Pool > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 2018/0322395 A1 - Document ID: US20180322395A1 - Status: Published application - Kind: A1 - Application number: 16/038,074 - Filed: 2018-07-17 - Published: 2018-11-08 - Earliest claimed: 2014-01-28 - Assignees: Sentient Technologies (Barbados) Limited - Applicants: Sentient Technologies (Barbados) Limited - Inventors: Hormoz Shahrzad; Babak Hodjat - Web page: https://patentgrep.com/patent/US20180322395A1 ## CPC - G06N 3/126: Evolutionary algorithms, e.g. genetic algorithms or genetic programming - G06N 5/025: Extracting rules from data - G06N 7/02: using fuzzy logic (computing arrangements based on biological models G06N3/00; computing arrangements using knowledge-based models G06N5/00) - G06N 20/10: using kernel methods, e.g. support vector machines [SVM] - G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS ## Abstract Roughly described, in an evolutionary technique for finding optimal solutions to a provided problem, a computer system uses a grouping algorithm that is better able to find diverse and optimum solutions in data mining environment with multiple solution landscapes and a plurality of candidate individuals. Each candidate individual identifies with a potential solution, and is associated with a testing experience level and one or more partition tags. Each candidate individual is assigned into one of a plurality of competition groups in dependence upon the individual's testing experience level and partition tag. During competition among candidate individuals, a candidate individual can only replace another candidate individual if both the candidate individuals have a common partition tag and are in the same competition group. A candidate individual cannot replace another candidate individual if they have different partition tags or are in different competition groups. ## Claims ### Claim 1 A computer-implemented method for discovering solutions to a provided problem, comprising the steps of: providing a computer system having a memory having a candidate pool database identifying a pool of candidate individuals, each candidate individual identifying a potential solution to the problem, each candidate individual further having associated therewith a respective fitness estimate, a respective testing experience level, and a respective value for a partition tag; a computer system testing individuals from the pool of candidate individuals against training data; a computer system updating the fitness estimates for tested ones of the individuals, in dependence upon the results of such testing; a computer system assigning each candidate individual in the pool of candidate individuals into one of a plurality of competition groups in dependence upon the testing experience level and the partition tag value of the candidate individual; and a computer system selecting individuals for discarding from the pool of candidate individuals in dependence upon their assigned competition group. ### Claim 2 (depends on 1) The method of claim 1, wherein each of the candidate individuals is associated with values for a plurality of partition tags, and wherein the step of assigning comprises assigning each of the candidate individuals to one of the competition groups in dependence upon a combination of the plurality of partition tag values associated with the candidate individual. ### Claim 3 (depends on 1) The method of claim 1, wherein the each of the competition groups is associated with a range of partition tag values, and wherein assigning each candidate individual into a competition group comprises assigning the candidate individual to a competition group associated with a range of partition tag values that includes the partition tag value of the candidate individual. ### Claim 4 (depends on 1) The method of claim 1, wherein the competition group to which each individual is assigned is further dependent upon the individual's testing experience level. ### Claim 5 (depends on 1) The method of claim 1, wherein the partition tag of the candidate individual is a property of the individual. ### Claim 6 (depends on 1) The method of claim 1, wherein the partition tag of the candidate individual is an ancestry count, the partition tag value indicating a number of procreation events in the ancestry of the individual. ### Claim 7 (depends on 1) The method of claim 1, wherein the partition tag value of the candidate individual is a predefined measure of diversity of the candidate individual relative to other individuals in the candidate individual pool. ### Claim 8 (depends on 1) The method of claim 1, wherein the problem includes a plurality of sub-problems, and wherein the partition tag value identifies one of the sub-problems addressed by the candidate individual. ### Claim 9 (depends on 1) The method of claim 1, further comprising a step of: generating a new candidate individual from one or more parent individuals, wherein the partitions tag of the new candidate individual is dependent upon the partition tags of the new candidate's parent individuals. ### Claim 10 (depends on 9) The method of claim 9, further comprising steps of: including the new candidate individual to the candidate pool database; and repeating the steps of testing, updating, assigning and selecting. ### Claim 11 (depends on 1) The method of claim 1, wherein selecting individuals for discarding from the pool of candidate individuals comprises: discarding from a particular one of the competition groups the least fit individual in the particular competition group as indicated by the fitness estimates. ### Claim 12 (depends on 1) The method of claim 1, wherein selecting individuals for discarding from the pool of candidate individuals comprises: discarding from each particular one of the competition groups in a subset of the competition groups the least fit individual in the particular competition group as indicated by the fitness estimates. ### Claim 13 (depends on 1) The method of claim 1, wherein selecting individuals for discarding is performed further in dependence upon other predetermined criteria. ### Claim 14 A system for discovering solutions to a provided problem, comprising: a memory storing a candidate pool database identifying a pool of candidate individuals, each candidate individual identifying a potential solution to the problem, each candidate individual further having associated therewith a respective fitness estimate, a respective testing experience level, and a respective value for a partition tag; and a processing unit which: tests individuals from the pool of candidate individuals against training data, updates the fitness estimates for tested ones of the individuals, in dependence upon the results of such testing, assigns each candidate individual in the pool of candidate individuals into one of a plurality of competition groups in dependence upon the testing experience level and the partition tag value of the candidate individual, and selects individuals for discarding from the pool of candidate individuals in dependence upon their assigned competition group. ### Claim 15 (depends on 14) The system of claim 14, wherein each of the candidate individuals is associated with values for a plurality of partition tags, and wherein the step of assigning comprises assigning each of the candidate individuals to one of the competition groups in dependence upon a combination of the plurality of partition tag values associated with the candidate individual. ### Claim 16 (depends on 14) The system of claim 14, wherein the each of the competition groups is associated with a range of partition tag values, and wherein assigning each candidate individual into a competition group comprises assigning the candidate individual to a competition group associated with a range of partition tag values that includes the partition tag value of the candidate individual. ### Claim 17 (depends on 14) The system of claim 14, wherein the competition group to which each individual is assigned is further dependent upon the individual's testing experience level. ### Claim 18 (depends on 14) The system of claim 14, wherein the partition tag of the candidate individual is a property of the individual. ### Claim 19 (depends on 14) The system of claim 14, wherein the partition tag of the candidate individual is an ancestry count, the partition tag value indicating a number of procreation events in the ancestry of the individual. ### Claim 20 (depends on 14) The system of claim 14, wherein the partition tag value of the candidate individual is a predefined measure of diversity of the candidate individual relative to other individuals in the candidate individual pool. ### Claim 21 (depends on 14) The system of claim 14, wherein the problem includes a plurality of sub-problems, and wherein the partition tag value identifies one of the sub-problems addressed by the candidate individual. ### Claim 22 (depends on 14) The system of claim 14, wherein the processing unit generates a new candidate individual from one or more parent individuals, and wherein the partitions tag of the new candidate individual is dependent upon the partition tags of the new candidate's parent individuals. ### Claim 23 (depends on 22) The system of claim 22, wherein the processing unit: includes the new candidate individual to the candidate pool database; and repeats the steps of testing, updating, assigning and selecting. ### Claim 24 (depends on 14) The system of claim 14, wherein selecting individuals for discarding from the pool of candidate individuals comprises: discarding from a particular one of the competition groups the least fit individual in the particular competition group as indicated by the fitness estimates. ### Claim 25 (depends on 14) The system of claim 14, wherein selecting individuals for discarding from the pool of candidate individuals comprises: discarding from each particular one of the competition groups in a subset of the competition groups the least fit individual in the particular competition group as indicated by the fitness estimates. ### Claim 26 (depends on 14) The system of claim 14, wherein selecting individuals for discarding is performed further in dependence upon other predetermined criteria. ### Claim 27 A non-transitory computer readable medium having stored thereon a plurality of instructions for discovering solutions to a provided problem, the plurality of instructions when executed by a processing unit causing the processing unit to: provide in a computer system memory a candidate pool database identifying a pool of candidate individuals, each candidate individual identifying a potential solution to the problem, each candidate individual further having associated therewith a respective fitness estimate, a respective testing experience level, and a respective value for a partition tag; test individuals from the pool of candidate individuals against training data; update the fitness estimates for tested ones of the individuals, in dependence upon the results of such testing; assign each candidate individual in the pool of candidate individuals into one of a plurality of competition groups in dependence upon the testing experience level and the partition tag value of the candidate individual; and select individuals for discarding from the pool of candidate individuals in dependence upon their assigned competition group. ### Claim 28 A computer-implemented data mining system for discovering solutions to a provided problem, comprising: a memory having a candidate pool database identifying a pool of candidate individuals, each candidate individual identifying a potential solution to the problem, each candidate individual further having associated therewith a respective fitness estimate, a respective testing experience level, and a respective value for a partition tag; means for testing individuals from the pool of candidate individuals against training data; means for updating the fitness estimates for tested ones of the individuals, in dependence upon the results of such testing; means for assigning each candidate individual in the pool of candidate individuals into one of a plurality of competition groups in dependence upon the testing experience level and the partition tag value of the candidate individual; and means for selecting individuals for discarding from the pool of candidate individuals in dependence upon their assigned competition group.