Artificial intelligence umbilical cord stem cell cultivation parameter control method, system and device

Abstract

The invention relates to an artificial intelligence umbilical cord stem cell culture parameter control method, system and device, and belongs to the technical field of stem cell culture. The method comprises the following steps: assigning a culture intervention medium of umbilical cord stem cells to obtain a plurality of medium distribution coordinates; obtaining a first-stage cultivation sample set and a second-stage cultivation sample set, and calculating a plurality of character deviation coefficients with the cultivation character index expected value; and optimizing the plurality of medium distribution coordinates to obtain target medium distribution coordinates, and sending the target medium distribution coordinates to an umbilical cord stem cell culture database. The technical problems that in the prior art, umbilical cord stem cell culture parameters depend on artificial experience configuration, so that ideal effects can be achieved only through multiple tests, and the response efficiency is poor are solved, automatic optimization configuration is carried out on the culture intervention medium through the artificial intelligence algorithm, the test frequency is reduced, and the test efficiency is improved. The parameter optimization efficiency and the culture success rate of the umbilical cord stem cells are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of stem cell cultivation, and particularly to an artificial intelligence-based method, system and device for controlling the cultivation parameters of umbilical cord stem cells. Background Art

[0002] Due to characteristics such as multi-directional differentiation potential and low immunogenicity, umbilical cord stem cells have broad application prospects in the fields of regenerative medicine and cell therapy. However, during the in vitro cultivation process of umbilical cord stem cells, the control of cultivation parameters directly affects the trait indicators of stem cells, such as key indicators like morphology, marker expression, proliferation ability, population doubling time, cell survival rate, chromosome stability, and microbial contamination control.

[0003] In the prior art, the control of umbilical cord stem cell cultivation parameters mainly relies on manual experience configuration, such as the selection of culture medium components, the setting of culture conditions, the determination of cell seeding density and passage timing, the culture surface treatment method, and the quality control system. This traditional parameter control method has obvious limitations: First, the parameter configuration overly relies on the experience of experimental personnel, resulting in low standardization; second, to obtain ideal cultivation effects, multiple repeated experiments and parameter adjustments are usually required, which not only consume time and effort but also have high cultivation costs; at the same time, the method of manual multiple experimental adjustments leads to poor response efficiency in the parameter optimization process and is difficult to quickly adapt to the individual differences of different batches of umbilical cord stem cells. Summary of the Invention

[0004] Aiming at the technical problem in the prior art that the cultivation parameters of umbilical cord stem cells rely on manual experience configuration, resulting in the need for multiple experiments to possibly obtain ideal effects and poor response efficiency, the present invention provides an artificial intelligence-based method, system and device for controlling the cultivation parameters of umbilical cord stem cells to solve this problem.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: In a first aspect, the present invention provides a method for controlling the cultivation parameters of umbilical cord stem cells by artificial intelligence, including: assigning values to the cultivation intervention media of umbilical cord stem cells to obtain a number of media distribution coordinates; traversing the number of media distribution coordinates and collecting adjacent samples of trait indicators through experiments and literature to obtain a first-level cultivation sample set, where the first-level cultivation samples include a first-level media distribution coordinate set; traversing the first-level media distribution coordinate set and collecting adjacent samples of trait indicators through experiments and literature to obtain a second-level cultivation sample set; performing a central value fitting on the number of media distribution coordinates based on the trait indicators of the first-level cultivation sample set and the second-level cultivation sample set to obtain a number of predicted values of cultivation trait indicators, and calculating a number of trait deviation coefficients from the expected values of the cultivation trait indicators; optimizing the number of media distribution coordinates according to the number of trait deviation coefficients to obtain target media distribution coordinates and sending them to the umbilical cord stem cell cultivation database.

[0006] Optionally, assigning values to the cultivation intervention media of umbilical cord stem cells to obtain a number of media distribution coordinates includes: the cultivation intervention media includes discrete intervention media and continuous intervention media; randomly assigning values based on the discrete intervention media constraint value set to obtain a first assignment result; randomly assigning values based on the continuous intervention media constraint interval to obtain a second assignment result; combining the first assignment result and the second assignment result to obtain a first media distribution coordinate and adding it to the number of media distribution coordinates.

[0007] Optionally, combining the first assignment result and the second assignment result to obtain a first media distribution coordinate includes: obtaining a media distribution coordinate system and abnormal detection materials; combining the first assignment result and the second assignment result and placing them into the media distribution coordinate system to obtain an initial media distribution coordinate; performing abnormal detection on the initial media distribution coordinate through the abnormal detection materials to obtain a coordinate abnormality degree; when the coordinate abnormality degree is equal to 0, setting the initial media distribution coordinate as the first media distribution coordinate.

[0008] Optionally, perform anomaly detection on the initial medium distribution coordinates using the anomaly detection material to obtain the coordinate anomaly degree, including: Step 1: The anomaly detection material includes several anomaly distribution coordinates, where any one of the several anomaly distribution coordinates includes at least two groups, and the several anomaly distribution coordinates are obtained by integrating and networking the collected cultivation anomaly parameters; Step 2: Add the initial medium distribution coordinates to the several anomaly distribution coordinates to obtain an anomaly detection distribution coordinate set; Step 3: Perform plane cutting and bisection on the anomaly detection distribution coordinate set to obtain a unilateral anomaly detection distribution coordinate set with the initial medium distribution coordinates, which is set as the first-level anomaly detection distribution coordinate set; Step 4: When the number of the first-level anomaly detection distribution coordinate sets is not equal to 1 and the bisection times are less than or equal to the preset times, return to Step 3 based on the first-level anomaly detection distribution coordinate set to execute the loop; Step 5: When the number of the first-level anomaly detection distribution coordinate sets is equal to 1, configure the coordinate anomaly degree to 0; when the number of the first-level anomaly detection distribution coordinate sets is not equal to 1 and the bisection times are greater than the preset times, configure the coordinate anomaly degree to 1.

[0009] Optionally, traverse the several medium distribution coordinates, combine experiments and literature to collect adjacent samples of trait indicators, and obtain a first-level cultivation sample set, including: obtaining the attribute of the trait indicator to be analyzed; based on the attribute of the trait indicator to be analyzed, perform cultivation weight distribution on the intervention medium attribute set through grey relational analysis to obtain the intervention medium weight distribution; based on the intervention medium weight distribution, combine the Euclidean distance threshold, traverse the several medium distribution coordinates to construct several adjacent constraint distances, and combine experiments and literature to collect the first-level first cultivation sample set belonging to the trait indicator to be analyzed, and add it to the first-level cultivation sample set.

[0010] Optionally, before combining the Euclidean distance threshold based on the intervention medium weight distribution, traversing the several medium distribution coordinates to construct several adjacent constraint distances, and combining experiments and literature to collect the first-level first cultivation sample set belonging to the trait indicator to be analyzed, it also includes: based on the intervention medium weight distribution, count the Euclidean distances of multiple historical intervention media when the fluctuation value of the attribute of the trait indicator to be analyzed is less than or equal to the fluctuation threshold of the trait indicator to be analyzed, where any one of the Euclidean distances of the multiple historical intervention media has a trigger frequency identifier; extract the minimum Euclidean distance with the trigger frequency identifier greater than or equal to the trigger frequency threshold from the Euclidean distances of the multiple historical intervention media, which is set as the Euclidean distance threshold.

[0011] Optionally, based on the trait indicators of the first-level cultivation sample set and the second-level cultivation sample set, perform a central value fitting on the several medium distribution coordinates to obtain several sets of predicted cultivation trait indicators, including: extracting the first-level cultivation sample set and the second-level cultivation sample set of the first trait indicator attribute; grouping the second-level cultivation sample set according to the first-level cultivation sample set to obtain multiple groups of second-level cultivation samples; traversing the multiple groups of second-level cultivation samples to calculate the central values of the trait indicators respectively, obtaining multiple central values of the second-level trait indicators; after calculating the mean values of the trait indicators between the multiple central values of the second-level trait indicators and the corresponding first-level cultivation sample sets one by one, perform a central value calculation of the trait indicators again to obtain the predicted value of the first cultivation trait indicator, and add it to the first group of predicted cultivation trait indicators and add it to the several sets of predicted cultivation trait indicators.

[0012] Optionally, optimize the several medium distribution coordinates according to the several trait deviation coefficients to obtain the target medium distribution coordinates and send them to the umbilical cord stem cell cultivation database, including: configuring the coordinate search rule: sequentially extract the first medium distribution coordinate, the second medium distribution coordinate, and the third medium distribution coordinate from the several medium distribution coordinates according to the several trait deviation coefficients from small to large; calculate the coordinate mean values of the first medium distribution coordinate, the second medium distribution coordinate, and the third medium distribution coordinate to obtain the search reference coordinate; perform coordinate search on the several medium distribution coordinates based on the search reference coordinate to obtain the updated medium distribution coordinates and perform optimization iteration.

[0013] In a second aspect, the present invention provides an artificial intelligence umbilical cord stem cell cultivation parameter control system, including: a medium assignment module for assigning cultivation intervention media to umbilical cord stem cells to obtain several medium distribution coordinates; a sample collection module for traversing the several medium distribution coordinates to collect adjacent samples of trait indicators in combination with experiments and literature to obtain a first-level cultivation sample set, where the first-level cultivation samples include a first-level medium distribution coordinate set; an adjacent collection module for traversing the first-level medium distribution coordinate set to collect adjacent samples of trait indicators in combination with experiments and literature to obtain a second-level cultivation sample set; a fitting prediction module for performing central value fitting on the several medium distribution coordinates based on the trait indicators of the first-level cultivation sample set and the second-level cultivation sample set to obtain several sets of predicted cultivation trait indicators, and calculating several trait deviation coefficients from the expected values of the cultivation trait indicators; an optimization module for optimizing the several medium distribution coordinates according to the several trait deviation coefficients to obtain the target medium distribution coordinates and send them to the umbilical cord stem cell cultivation database.

[0014] In a third aspect, the present invention provides a control device for umbilical cord stem cell culture parameters of artificial intelligence, including a control system for umbilical cord stem cell culture parameters of artificial intelligence.

[0015] The beneficial effects of the present invention are as follows: Assign values to the culture intervention media for umbilical cord stem cells to obtain several media distribution coordinates. By parameterizing the culture intervention media parameters such as culture medium components, culture conditions, cell seeding and passage, culture surface treatment, pollution control, and quality control, they are transformed into quantifiable media distribution coordinates, laying a foundation for subsequent intelligent optimization; traverse several media distribution coordinates and collect adjacent samples of trait indicators through experiments and literature to obtain a first-level culture sample set. Among them, the first-level culture samples include a first-level media distribution coordinate set. By systematically collecting the trait indicator data of umbilical cord stem cell culture under different media distribution coordinates, including key indicators such as morphology, marker detection values, and proliferation rates, an initial data set is formed to provide data support for subsequent model establishment; traverse the first-level media distribution coordinate set and collect adjacent samples of trait indicators through experiments and literature to obtain a second-level culture sample set. By further expanding the data collection range on the basis of the first-level samples, a richer correspondence between culture parameters and trait indicators is obtained, enhancing the representativeness and comprehensiveness of the data and improving the accuracy of the subsequent prediction model; based on the trait indicators of the first-level culture sample set and the second-level culture sample set, perform median value fitting on several media distribution coordinates to obtain several sets of predicted values of culture trait indicators, and calculate several trait deviation coefficients from the expected values of the culture trait indicators. By calculating the deviation coefficients between the predicted values and the expected values, the quality of the current parameter configuration is quantified, providing a basis for subsequent parameter optimization; optimize several media distribution coordinates according to several trait deviation coefficients, obtain the target media distribution coordinates and send them to the umbilical cord stem cell culture database. By performing intelligent optimization based on the deviation coefficients, the optimal media distribution coordinates are determined and stored in the database to achieve automatic optimization configuration of umbilical cord stem cell culture parameters.

[0016] Through the above technical solutions, the technical problem in the prior art that the umbilical cord stem cell culture parameters rely on manual experience configuration, resulting in the need for multiple experiments to possibly obtain ideal results and poor response efficiency, is solved. The artificial intelligence algorithm is used to systematically analyze and automatically optimize the culture intervention media, reducing the number of experiments, improving the parameter optimization efficiency and the success rate of umbilical cord stem cell culture, and enhancing the intelligent level and culture effect of the umbilical cord stem cell culture process. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic flowchart of a control method for umbilical cord stem cell culture parameters of artificial intelligence provided by the present invention; Figure 2Schematic diagram of a parameter control system for culturing umbilical cord stem cells with artificial intelligence provided by the present invention.

[0018] In the drawings, the components represented by the reference numerals are as follows: Medium assignment module 11, sample collection module 12, adjacent collection module 13, fitting prediction module 14, optimization module 15. Detailed implementation manners

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative efforts belong to the protection scope of the present invention.

[0020] In the description of the present invention, the terms "first" and "second" are only used for descriptive purposes, and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include one or more of the described features. In the description of the present invention, "a plurality of" means two or more, unless otherwise specifically defined.

[0021] In the description of the present invention, the term "for example" is used to mean "serving as an example, illustration, or explanation". Any embodiment described as "for example" in the present invention is not necessarily construed as being more preferred or more advantageous than other embodiments. The following description is given to enable any person skilled in the art to make and use the present invention. In the following description, details are set forth for the purpose of explanation. It should be understood that those skilled in the art can recognize that the present invention can be implemented without these specific details. In other instances, well-known structures and processes are not described in detail to avoid unnecessary details from obscuring the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0022] Embodiment 1, as Figure 1 shown, the embodiment of the present invention provides a method for controlling the parameters of culturing umbilical cord stem cells with artificial intelligence, including: S100: Assign values to the culture intervention medium of umbilical cord stem cells to obtain a number of medium distribution coordinates.

[0023] Specifically, assign values to the cultivation intervention media of umbilical cord stem cells to obtain several media distribution coordinates, thereby setting initial values for various intervention parameters in the cultivation of umbilical cord stem cells and mapping them to a multi-dimensional coordinate space. Among them, the cultivation intervention media of umbilical cord stem cells refer to various parameters that affect the cultivation process of umbilical cord stem cells, including aspects such as culture medium components, culture conditions, cell seeding and passage, culture surface treatment, pollution control, and quality control.

[0024] First, within a preset parameter range, perform multiple different value assignment operations on each parameter of the cultivation intervention media. For example, assign values to the culture temperature within the range of 35 - 38 °C, and assign values to the pH value within the range of 7.2 - 7.6, etc. Each value assignment operation will generate a complete set of cultivation parameter combinations. After performing multiple value assignments on the cultivation intervention media of umbilical cord stem cells, specific numerical values are given to each parameter of several sets of cultivation intervention media. Then, map these parameters to a multi-dimensional parameter space to form several media distribution coordinates. Each media distribution coordinate represents a specific set of cultivation intervention media parameter combinations, and these coordinate points will serve as the basic data for subsequent analysis and optimization, laying a data foundation for subsequent parameter optimization based on artificial intelligence.

[0025] S200: Traverse the several media distribution coordinates, collect adjacent samples of trait indicators through experiments and literature, and obtain a first-level cultivation sample set. Among them, the first-level cultivation samples include a first-level media distribution coordinate set.

[0026] Specifically, after obtaining several media distribution coordinates, perform a traversal operation on these media distribution coordinates, and combine actual experimental data and research results in relevant literature to collect trait indicator data related to each media distribution coordinate. Among them, trait indicators refer to various biological characteristics of umbilical cord stem cells, including indicators such as cell morphology, marker detection values, proliferation rate, population doubling time, cell survival rate, and chromosomal abnormality values.

[0027] For each media distribution coordinate, collect relevant trait indicator data by analyzing experimental results and literature with similar or close parameters to it, and achieve adjacent sample collection. Among them, adjacent samples refer to sample points that are relatively close to the current media distribution coordinate in the multi-dimensional parameter space, and the trait indicator data of these sample points can provide a reference for the trait performance of the current media distribution coordinate.

[0028] Through adjacent sample collection, corresponding trait indicator data can be established for each media distribution coordinate to form a first-level cultivation sample set. The first-level cultivation sample set contains a first-level media distribution coordinate set, that is, the screened and verified media distribution coordinates and their corresponding trait indicator data, providing preliminary data support for subsequent parameter optimization.

[0029] S300: Traverse the set of first-level medium distribution coordinates, collect adjacent samples of trait indicators by combining experiments and literature, and obtain a set of second-level cultivation samples.

[0030] Specifically, after obtaining the set of first-level cultivation samples and the set of first-level medium distribution coordinates it contains, continue with data collection. Specifically, traverse the set of first-level medium distribution coordinates, collect adjacent samples of trait indicators by combining experiments and literature, and obtain a set of second-level cultivation samples.

[0031] The method for collecting and obtaining the set of second-level cultivation samples is similar to the method for collecting adjacent samples for obtaining the set of first-level cultivation samples, but the processing object changes from the initial several medium distribution coordinates to the set of first-level medium distribution coordinates after the first screening. This hierarchical collection strategy enables more refined data collection within a more valuable parameter space region.

[0032] Specifically, for each coordinate point in the set of first-level medium distribution coordinates, collect the trait indicator data related to this coordinate point again by combining experimental data and literature, thereby forming a set of second-level cultivation samples. The set of second-level cultivation samples is a sample set collected further on the basis of the first-level collection. The set of second-level cultivation samples contains more trait indicator data for umbilical cord stem cell cultivation, providing a data basis for subsequent parameter optimization.

[0033] Through the collection of the set of first-level cultivation samples and the set of second-level cultivation samples, a hierarchical cultivation sample database is constructed, providing a strong guarantee for the accuracy and effectiveness of parameter optimization.

[0034] S400: Based on the trait indicators of the set of first-level cultivation samples and the set of second-level cultivation samples, perform a central value fitting on the several medium distribution coordinates, obtain several sets of predicted values of cultivation trait indicators, and calculate several trait deviation coefficients from the expected values of cultivation trait indicators.

[0035] Specifically, after obtaining the set of first-level cultivation samples and the set of second-level cultivation samples, first, perform a central value fitting calculation on the initial several medium distribution coordinates by using the trait indicator data in the previously collected set of first-level cultivation samples and the set of second-level cultivation samples. Central value fitting refers to inferring the trait indicator values that umbilical cord stem cells may exhibit under specific medium distribution coordinates through statistical methods based on the sample data in the existing set of first-level cultivation samples and the set of second-level cultivation samples. By performing a central value fitting on several medium distribution coordinates, the data distribution characteristics in the parameter space are considered, and the trait performance of the current parameter combination that has not been actually verified is accurately predicted.

[0036] By means of centralized value fitting, corresponding predicted values of cultivation trait indicators are generated for each medium distribution coordinate, forming several groups of predicted values of cultivation trait indicators. Then, these predicted values are compared and analyzed with the expected values of the cultivation trait indicators set in advance, and several trait deviation coefficients are calculated. The trait deviation coefficient reflects the degree of difference between the predicted value and the expected value, and is an important indicator for measuring whether a specific parameter combination meets the cultivation goal. By calculating several trait deviation coefficients, it is possible to evaluate the potential effects of the parameter combinations corresponding to each medium distribution coordinate without conducting a large number of actual experiments, providing a quantitative evaluation basis for subsequent parameter optimization, thereby improving the efficiency and accuracy of umbilical cord stem cell cultivation parameter optimization.

[0037] S500: Optimize the several medium distribution coordinates according to the several trait deviation coefficients, and obtain the target medium distribution coordinates to be sent to the umbilical cord stem cell cultivation database.

[0038] Specifically, after calculating several trait deviation coefficients, these coefficients are used as evaluation indicators to optimize and screen the initial several medium distribution coordinates. Among them, the trait deviation coefficient reflects the degree of difference between the predicted value of the cultivation trait indicator and the expected value of the cultivation trait indicator under a specific medium distribution coordinate. The smaller the trait deviation coefficient, the closer the parameter combination corresponding to the medium distribution coordinate is to the ideal cultivation effect.

[0039] By analyzing the distribution characteristics of several trait deviation coefficients, optimize and adjust the several medium distribution coordinates to achieve optimization of the several medium distribution coordinates. For example, construct a search space based on the medium distribution coordinates with smaller trait deviation coefficients, and generate new medium distribution coordinates that may have smaller trait deviation coefficients through iterative calculation. When the trait deviation coefficient of a certain medium distribution coordinate reaches a preset threshold (such as less than 0.05), it is determined as the target medium distribution coordinate. After obtaining the target medium distribution coordinates, send them to the umbilical cord stem cell cultivation database for storage and management. Among them, the target medium distribution coordinates represent an intelligent optimized parameter combination for umbilical cord stem cell cultivation, which can be directly applied to the actual cultivation process, providing a basis for improving cultivation efficiency and trait stability.

[0040] By optimizing the several medium distribution coordinates to obtain the target medium distribution coordinates, realizing the intelligent optimization of umbilical cord stem cell cultivation parameters. Compared with the traditional manual trial-and-error method, it improves the efficiency and accuracy of parameter optimization, reduces the consumption of experimental resources, and provides technical support for the large-scale cultivation of umbilical cord stem cells.

[0041] Furthermore, assign values to the cultivation intervention media of umbilical cord stem cells to obtain several medium distribution coordinates, including: S110: The cultivation intervention media include discrete intervention media and continuous intervention media; S120: Randomly assign values based on the set of discrete intervention medium constraint values to obtain the first assignment result; S130: Randomly assign values based on the continuous intervention medium constraint interval to obtain the second assignment result; S140: Combine the first assignment result and the second assignment result to obtain the first medium distribution coordinate and add it to the several medium distribution coordinates.

[0042] Specifically, the cultivation intervention media are divided into two categories, including discrete intervention media and continuous intervention media. Among them, discrete intervention media refer to parameters that can only take specific discrete values, such as culture medium type, additive type, culture dish material, etc.; while continuous intervention media refer to parameters that can continuously take values within a certain interval, such as culture temperature, pH value, carbon dioxide concentration, oxygen concentration, etc.

[0043] For discrete intervention media, randomly assign values based on their set of constraint values to obtain the first assignment result. Specifically, for each discrete intervention medium, randomly select a value from its set of optional discrete intervention medium constraint values as the assignment result. For example, the possible set of constraint values for the culture medium type is {DMEM, RPMI - 1640, α - MEM}, and a specific type is randomly selected as part of the first assignment result. After completing a single random assignment for all discrete intervention media, the first assignment result is obtained.

[0044] For continuous intervention media, randomly assign values based on their constraint interval to obtain the second assignment result. Specifically, for each continuous intervention medium, randomly generate a value within its preset value range. For example, the constraint interval for the culture temperature may be [35°C, 38°C], and a temperature value is randomly selected within this interval as part of the second assignment result. After completing a single random assignment for all continuous intervention media, the second assignment result is obtained.

[0045] After performing a single assignment on the cultivation intervention media, the first assignment result for discrete intervention media and the second assignment result for continuous intervention media are obtained. Combine the obtained first assignment result and the second assignment result to form a complete parameter combination, that is, the first medium distribution coordinate, and add it to the several medium distribution coordinates. By repeating the random assignment process multiple times, different several medium distribution coordinates are generated, providing a data basis for subsequent optimization analysis.

[0046] By classifying and assigning values to different types of cultivation intervention media, different characteristics of the cultivation intervention media parameters are fully considered, making the value assignment process more in line with the actual situation. At the same time, diverse parameter combinations are generated randomly, enhancing the coverage of the parameter space and facilitating the discovery of the optimal cultivation parameter combination.

[0047] Further, combining the first assignment result and the second assignment result to obtain the first medium distribution coordinate, including: S141: Obtain the medium distribution coordinate system and the abnormal detection material; S142: Combine the first assignment result and the second assignment result, and place them into the medium distribution coordinate system to obtain the initial medium distribution coordinate; S143: Perform abnormal detection on the initial medium distribution coordinate through the abnormal detection material to obtain the coordinate abnormality degree; S144: When the coordinate abnormality degree is equal to 0, set the initial medium distribution coordinate as the first medium distribution coordinate.

[0048] Specifically, first, obtain the medium distribution coordinate system and the abnormal detection material. Among them, the medium distribution coordinate is a multi-dimensional coordinate space used to represent and locate various cultivation intervention medium parameter combinations; the abnormal detection material contains information on known abnormal parameter combinations and is used to screen whether there are potential problems with newly generated parameter combinations.

[0049] When combining the first assignment result and the second assignment result, first, combine the obtained first assignment result (the assignment of discrete intervention media) and the obtained second assignment result (the assignment of continuous intervention media) to form a complete parameter combination, and place it into the medium distribution coordinate system to obtain the initial medium distribution coordinate. Then, use the abnormal detection material to perform abnormal detection on the initial medium distribution coordinate, and calculate the coordinate abnormality degree. The coordinate abnormality degree is an index to measure whether the parameter combination is abnormal, and this index is determined by comparing the similarity or proximity of the initial medium distribution coordinate with the known abnormal parameter combinations. Among them, the abnormal parameter combination is the parameter setting that has been proven to cause cultivation failure or extremely poor effects in the experiment.

[0050] Then, make a judgment according to the abnormal detection result. When the coordinate abnormality degree is equal to 0, it means that there is no obvious abnormality in the initial medium distribution coordinate and it can be used as a valid parameter combination. At this time, set the initial medium distribution coordinate as the first medium distribution coordinate and add it to several medium distribution coordinates. If the coordinate abnormality degree is not equal to 0, it means that there may be problems with this parameter combination and the random assignment process needs to be carried out again to generate new parameter combinations.

[0051] Through the anomaly detection mechanism, parameter combinations that may lead to cultivation failure can be effectively filtered out, improving the efficiency and accuracy of subsequent optimization analysis, reducing invalid experiments, and saving research resources.

[0052] Furthermore, the initial medium distribution coordinates are subjected to anomaly detection by the anomaly detection material to obtain the coordinate anomaly degree, including: Step 1: The anomaly detection material includes a number of abnormal distribution coordinates. Among them, any one of the abnormal distribution coordinates of the number of abnormal distribution coordinates includes at least two groups, and the number of abnormal distribution coordinates is obtained by integrating and networking the collected cultivation abnormal parameters; Step 2: Add the initial medium distribution coordinates to the number of abnormal distribution coordinates to obtain an anomaly detection distribution coordinate set; Step 3: Perform plane cutting and bisection on the anomaly detection distribution coordinate set to obtain a unilateral anomaly detection distribution coordinate set with the initial medium distribution coordinates, which is set as the first-level anomaly detection distribution coordinate set; Step 4: When the number of the first-level anomaly detection distribution coordinate sets is not equal to 1 and the number of bisections is less than or equal to the preset number of times, return to Step 3 based on the first-level anomaly detection distribution coordinate set to execute the loop; Step 5: When the number of the first-level anomaly detection distribution coordinate sets is equal to 1, configure the coordinate anomaly degree to 0; when the number of the first-level anomaly detection distribution coordinate sets is not equal to 1 and the number of bisections is greater than the preset number of times, configure the coordinate anomaly degree to 1.

[0053] Specifically, the anomaly detection material includes a number of abnormal distribution coordinates, which are obtained by integrating and networking the collected cultivation abnormal parameters and represent parameter combinations that are known to cause cultivation failure or poor effects. It should be noted that any one of the number of abnormal distribution coordinates includes at least two groups, which means that each abnormal situation has multiple groups of parameter data as support, ensuring the reliability of abnormal judgment.

[0054] When performing anomaly detection on the initial medium distribution coordinates by detecting materials through anomalies to obtain the coordinate anomaly degree, first, add the initial medium distribution coordinates to be detected into several anomaly distribution coordinates to form a complete set of anomaly detection distribution coordinates, so as to compare and analyze the parameter combination to be detected and the known anomaly parameter combination in the same coordinate space. Then, perform a plane cutting and bisection operation on the set of anomaly detection distribution coordinates. Among them, plane cutting and bisection means selecting a dimension as the cutting dimension in the multi-dimensional parameter space, and then finding a cutting point on this dimension (for example, the median or average value of all coordinate values on this dimension), and dividing the entire coordinate set into two subsets. Then select the subset that contains the initial medium distribution coordinates, which is called the one-sided anomaly detection distribution coordinate set, and set it as the first-level anomaly detection distribution coordinate set.

[0055] Subsequently, check whether the number of the first-level anomaly detection distribution coordinate sets is equal to 1, and whether the current bisection times are less than or equal to the preset times. Among them, the bisection times refer to the cumulative number of times of performing the plane cutting and bisection operation. Each time step three is executed, the bisection times increase by 1.

[0056] If the number of the first-level anomaly detection distribution coordinate sets is not equal to 1 (that is, it still contains other anomaly distribution coordinates), and the bisection times do not exceed the preset times, then return to step three based on the current first-level anomaly detection distribution coordinate set and continue to perform the plane cutting and bisection operation.

[0057] If the number of the first-level anomaly detection distribution coordinate sets is equal to 1, it means that after the bisection operation, the initial medium distribution coordinates have been separated from all anomaly distribution coordinates. At this time, configure the coordinate anomaly degree to 0, indicating that there is no anomaly at this coordinate point.

[0058] If the number of the first-level anomaly detection distribution coordinate sets is not equal to 1 (that is, it still contains other anomaly distribution coordinates), and the bisection times are greater than the preset times, it means that even after multiple bisections, the initial medium distribution coordinates are still in the same area as some anomaly distribution coordinates and cannot be separated. At this time, configure the coordinate anomaly degree to 1, indicating that there is an anomaly at this coordinate point.

[0059] Through anomaly detection based on spatial recursive segmentation, continuously try to separate the initial medium distribution coordinates to be measured from the known anomaly distribution coordinates. If successful separation can be achieved, it is considered that there is no anomaly in the initial medium distribution coordinates; if separation still cannot be achieved after the preset number of bisection operations, it is considered that the initial medium distribution coordinates are similar to some anomaly points and may have anomalies, so as to effectively identify the parameter combinations that may lead to cultivation failure and improve the accuracy and efficiency of parameter screening.

[0060] Further, traverse the several medium distribution coordinates, combine experiments and literature to collect adjacent samples of trait indicators, and obtain a first-level cultivation sample set, including: S210: Obtain the attribute of the trait indicator to be analyzed; S220: Based on the attribute of the trait indicator to be analyzed, perform cultivation weight distribution on the intervention medium attribute set through grey relational analysis to obtain the intervention medium weight distribution; S230: Based on the intervention medium weight distribution, combine the Euclidean distance threshold, traverse the several medium distribution coordinates to construct several adjacent constraint distances, combine experiments and literature to collect the first-level first cultivation sample set belonging to the trait indicator to be analyzed, and add it to the first-level cultivation sample set.

[0061] Specifically, in the process of obtaining the first-level cultivation sample set, first, obtain the attribute of the trait indicator to be analyzed. Among them, the attribute of the trait indicator to be analyzed refers to the characteristic parameters of umbilical cord stem cells that need to be focused on and optimized, such as indicators in aspects of cell morphology, marker detection values, proliferation rate, population doubling time, cell survival rate, chromosome abnormality values, etc. These trait indicator attributes are important bases for evaluating the cultivation quality and effect of umbilical cord stem cells. Then, based on the obtained attribute of the trait indicator to be analyzed, perform cultivation weight distribution on the intervention medium attribute set through grey relational analysis to obtain the intervention medium weight distribution. Through grey relational analysis, the influence degree of different intervention medium attributes on the trait indicator to be analyzed can be revealed. Through this analysis, the influence weight of each intervention medium parameter (such as temperature, pH value, culture medium components, etc.) on a specific trait indicator can be determined, forming the intervention medium weight distribution, which provides a weight basis for subsequent adjacent sample collection. Subsequently, based on the obtained intervention medium weight distribution, combine the Euclidean distance threshold, and traverse several medium distribution coordinates to construct several adjacent constraint distances. Specifically, for each medium distribution coordinate, calculate its weighted Euclidean distance from other coordinate points according to the intervention medium weight distribution. When this distance is less than the preset Euclidean distance threshold, it is considered that the two coordinate points have an adjacent relationship. Then, combine experimental data and literature materials to collect the trait indicator data of the sample points adjacent to the current medium distribution coordinate, form the first-level first cultivation sample set belonging to the trait indicator to be analyzed, and add it to the first-level cultivation sample set.

[0062] Through the method based on grey relational analysis and Euclidean distance calculation, the influence weight of different cultivation parameters on specific trait indicators can be determined, and targeted adjacent sample collection can be carried out based on this, improving the relevance and representativeness of the collected data, and providing a more accurate data basis for subsequent parameter optimization.

[0063] Further, based on the weight distribution of the intervention medium, in combination with the Euclidean distance threshold, traverse the coordinates of the several medium distributions to construct several adjacent constraint distances, and collect the first-level first cultivation sample set belonging to the trait index to be analyzed in combination with experiments and literature. This also includes: S240: Based on the weight distribution of the intervention medium, count the Euclidean distances of multiple historical intervention media when the fluctuation value of the trait index to be analyzed is less than or equal to the fluctuation threshold of the trait index to be analyzed. Among them, any one of the Euclidean distances of the multiple historical intervention media has a trigger frequency identifier. S250: Extract the minimum Euclidean distance with a trigger frequency identifier greater than or equal to the trigger frequency threshold from the Euclidean distances of the multiple historical intervention media, and set it as the Euclidean distance threshold.

[0064] Specifically, before traversing the coordinates of several medium distributions based on the weight distribution of the intervention medium, in combination with the Euclidean distance threshold, to construct several adjacent constraint distances, and collecting the first-level first cultivation sample set belonging to the trait index to be analyzed in combination with experiments and literature, it is necessary to reasonably determine the Euclidean distance threshold first.

[0065] First, based on the obtained weight distribution of the intervention medium, conduct statistical analysis on historical cultivation data. Specifically, count the Euclidean distances of multiple historical intervention media under the condition that the fluctuation value of the trait index to be analyzed is less than or equal to the fluctuation threshold of the trait index to be analyzed. Among them, the fluctuation value of the trait index to be analyzed refers to the change range of the trait index under similar cultivation conditions; the fluctuation threshold of the trait index to be analyzed is a preset acceptable fluctuation range. For the multiple historical intervention medium Euclidean distances obtained by statistics, each historical intervention medium Euclidean distance has a trigger frequency identifier, which records the frequency of occurrence of this Euclidean distance value in historical data or the frequency of resulting in similar trait performances. Then, from the obtained Euclidean distances of multiple historical intervention media, extract those Euclidean distance values with a trigger frequency identifier greater than or equal to the trigger frequency threshold, and select the minimum value from them and set it as the Euclidean distance threshold. The trigger frequency threshold is a preset frequency limit. Only when the trigger frequency of the Euclidean distance reaches or exceeds this threshold is this distance considered to have sufficient reliability. By selecting the minimum Euclidean distance that meets the trigger frequency requirement as the threshold, it not only ensures the correlation of adjacent samples but also avoids data redundancy caused by too large a collection range.

[0066] By analyzing the relationship between the parameter distance and trait fluctuation in historical data, combined with the statistics of the trigger frequency, find the optimal Euclidean distance threshold, provide a basis for accurate adjacent sample collection, so that the subsequent adjacent sample collection can cover enough similar sample points and maintain the collection efficiency.

[0067] Furthermore, based on the trait indicators of the first-level cultivation sample set and the second-level cultivation sample set, perform a central value fitting on the several medium distribution coordinates to obtain several sets of predicted values of cultivation trait indicators, including: S410: Extract the first-level cultivation sample set and the second-level cultivation sample set with the first trait indicator attribute; S420: Group the second-level cultivation sample set according to the first-level cultivation sample set to obtain multiple groups of second-level cultivation samples; S430: Traverse the multiple groups of second-level cultivation samples and calculate the central values of the trait indicators respectively to obtain multiple central values of the second-level trait indicators; S440: After calculating the mean value of the trait indicators between the multiple central values of the second-level trait indicators and the corresponding first-level cultivation sample set one by one, then calculate the central value of the trait indicators to obtain the predicted value of the first cultivation trait indicator, add it to the first group of predicted values of the cultivation trait indicators, and add it to the several sets of predicted values of the cultivation trait indicators.

[0068] Specifically, in the process of obtaining several sets of predicted values of cultivation trait indicators, first, extract the first-level cultivation sample set and the second-level cultivation sample set with the first trait indicator attribute. Among them, the first trait indicator attribute refers to the specific umbilical cord stem cell trait characteristics that need to be predicted, such as the expression of specific markers, proliferation rate, etc. Select the sample data related to this trait indicator attribute from the already obtained first-level cultivation sample set and the second-level cultivation sample set as the basis for subsequent analysis. Then, group the second-level cultivation sample set according to the first-level cultivation sample set to obtain multiple groups of second-level cultivation samples. Specifically, group based on the medium distribution coordinates in the first-level cultivation sample set, and group the samples in the second-level cultivation sample set associated with each first-level medium distribution coordinate into one group to form multiple groups of second-level cultivation samples.

[0069] Subsequently, traverse the multiple groups of second-level cultivation samples, calculate the central values of the trait indicators for each group of samples respectively to obtain multiple central values of the second-level trait indicators. Among them, the central value of the trait indicator refers to the value that can represent the overall trait performance of the group of samples calculated by statistical methods (such as median, mode, weighted average, etc.), so as to summarize the trait data of each group of second-level cultivation samples into a representative central value, and obtain multiple central values of the second-level trait indicators. Then, calculate the mean value of the trait indicators between the multiple obtained central values of the second-level trait indicators and their corresponding first-level cultivation sample set, that is, calculate the average value for each first-level sample and its corresponding central value of the second-level trait indicator. Then, calculate the central value of the trait indicators for these mean values to obtain the predicted value of the first cultivation trait indicator, and add it to the first group of predicted values of the cultivation trait indicators, and further add it to the several sets of predicted values of the cultivation trait indicators.

[0070] Through the grouping based on the medium distribution coordinates and the multi-level central value calculation method, the trait data of the primary and secondary culture samples can be effectively integrated, and the trait performance of umbilical cord stem cells under specific medium distribution coordinates can be accurately predicted, providing data support for parameter optimization.

[0071] Further, optimizing the several medium distribution coordinates according to the several trait deviation coefficients, and sending the target medium distribution coordinates to the umbilical cord stem cell culture database, including: S510: Configure the coordinate search rule: S520: Extract the first medium distribution coordinate, the second medium distribution coordinate, and the third medium distribution coordinate from the several medium distribution coordinates in sequence according to the ascending order of the several trait deviation coefficients; S530: Calculate the coordinate mean of the first medium distribution coordinate, the second medium distribution coordinate, and the third medium distribution coordinate to obtain the search reference coordinate; S540: Perform coordinate search on the several medium distribution coordinates based on the search reference coordinate, and obtain the updated medium distribution coordinates to execute the optimization iteration.

[0072] Specifically, when obtaining the target medium distribution coordinates, first, configure the coordinate search rule. The coordinate search rule refers to the strategy and method that defines how to search in the parameter space to find the optimal medium distribution coordinates, including the search range, search step size, termination conditions, etc. These rules provide an operation guide for the subsequent optimization process. Then, according to the calculated several trait deviation coefficients, sort them in ascending order of their values, and extract the corresponding first medium distribution coordinate, the second medium distribution coordinate, and the third medium distribution coordinate from the several medium distribution coordinates in sequence. These three coordinate points represent the three points with the smallest trait deviation coefficients in the current parameter space, that is, the three parameter combinations closest to the culture target.

[0073] Then, calculate the arithmetic mean of the parameter values of each dimension of the first medium distribution coordinate, the second medium distribution coordinate, and the third medium distribution coordinate to form a new coordinate point as the search reference coordinate. This method is based on the assumption that the optimal solution may be near the current optimal points, and determines the central area for the next search by taking the mean of the three optimal points. Subsequently, perform coordinate search on the several medium distribution coordinates based on the obtained search reference coordinate, and obtain the updated medium distribution coordinates to execute the optimization iteration. Specifically, in the parameter space around the search reference coordinate, generate new medium distribution coordinate points according to the configured search rule, calculate the trait deviation coefficients for these new points, and then repeat the process of S520 to S540 for iterative optimization until the termination condition is met.

[0074] Through the gradient search method based on the trait deviation coefficient, the optimal combination of umbilical cord stem cell culture parameters can be efficiently found in the parameter space. Through iterative optimization, continuously approaching the region with a smaller trait deviation coefficient, the target medium distribution coordinates are finally obtained and sent to the umbilical cord stem cell culture database for storage and management, providing a parameter basis for the actual culture process.

[0075] Example 2, as Figure 2 shown, based on the same inventive concept as the method for controlling umbilical cord stem cell culture parameters of artificial intelligence provided in Example 1, the embodiment of the present invention further provides a control system for umbilical cord stem cell culture parameters of artificial intelligence, including: A medium assignment module 11 for assigning values to the culture intervention media of umbilical cord stem cells to obtain a number of medium distribution coordinates; A sample collection module 12 for traversing the number of medium distribution coordinates to collect adjacent samples of trait indicators in combination with experiments and literature to obtain a first-level culture sample set, where the first-level culture samples include a first-level medium distribution coordinate set; An adjacent collection module 13 for traversing the first-level medium distribution coordinate set to collect adjacent samples of trait indicators in combination with experiments and literature to obtain a second-level culture sample set; A fitting prediction module 14 for performing central value fitting on the number of medium distribution coordinates based on the trait indicators of the first-level culture sample set and the second-level culture sample set to obtain several groups of predicted values of culture trait indicators, and calculating several trait deviation coefficients from the expected values of the culture trait indicators; An optimization module 15 for optimizing the number of medium distribution coordinates according to the several trait deviation coefficients to obtain the target medium distribution coordinates and sending them to the umbilical cord stem cell culture database.

[0076] Further, the execution steps of the medium assignment module 11 include: The culture intervention media include discrete intervention media and continuous intervention media; Randomly assign values based on the discrete intervention medium constraint value set to obtain a first assignment result; Randomly assign values based on the continuous intervention medium constraint interval to obtain a second assignment result; Combine the first assignment result and the second assignment result to obtain the first medium distribution coordinate and add it to the number of medium distribution coordinates.

[0077] Further, the execution steps of the medium assignment module 11 further include: Obtain the medium distribution coordinate system and the abnormal detection materials; Combine the first assignment result and the second assignment result, place them into the medium distribution coordinate system, and obtain the initial medium distribution coordinates. Use the abnormal detection material to perform abnormal detection on the initial medium distribution coordinates to obtain the coordinate abnormality degree. When the coordinate abnormality degree is equal to 0, set the initial medium distribution coordinates as the first medium distribution coordinates.

[0078] Furthermore, the execution steps of the medium assignment module 11 further include: Step 1: The abnormal detection material includes a number of abnormal distribution coordinates. Among them, any one of the abnormal distribution coordinates of the number of abnormal distribution coordinates includes at least two groups, and the number of abnormal distribution coordinates is obtained by integrating and networking to collect cultivation abnormal parameters. Step 2: Add the initial medium distribution coordinates to the number of abnormal distribution coordinates to obtain an abnormal detection distribution coordinate set. Step 3: Perform plane cutting and bisection on the abnormal detection distribution coordinate set to obtain a unilateral abnormal detection distribution coordinate set with the initial medium distribution coordinates, which is set as the first-level abnormal detection distribution coordinate set. Step 4: When the number of the first-level abnormal detection distribution coordinate sets is not equal to 1 and the bisection times are less than or equal to the preset times, return to Step 3 based on the first-level abnormal detection distribution coordinate sets to execute the loop. Step 5: When the number of the first-level abnormal detection distribution coordinate sets is equal to 1, configure the coordinate abnormality degree to 0; when the number of the first-level abnormal detection distribution coordinate sets is not equal to 1 and the bisection times are greater than the preset times, configure the coordinate abnormality degree to 1.

[0079] Furthermore, the execution steps of the sample collection module 12 include: Obtain the property of the trait index to be analyzed. Based on the property of the trait index to be analyzed, perform cultivation weight distribution on the intervention medium property set through grey relational analysis to obtain the intervention medium weight distribution. Based on the intervention medium weight distribution, combine the Euclidean distance threshold, traverse the number of medium distribution coordinates to construct a number of adjacent constraint distances, and combine experiments and literature to collect a first-level first cultivation sample set belonging to the trait index to be analyzed, and add it to the first-level cultivation sample set.

[0080] Furthermore, the execution steps of the sample collection module 12 further include: Based on the weight distribution of the intervention medium, statistically calculate the Euclidean distances of multiple historical intervention media when the fluctuation value of the trait index attribute to be analyzed is less than or equal to the fluctuation threshold of the trait index to be analyzed. Among them, any one of the Euclidean distances of the multiple historical intervention media has a trigger frequency identifier; Extract the minimum Euclidean distance with a trigger frequency identifier greater than or equal to the trigger frequency threshold from the Euclidean distances of the multiple historical intervention media, and set it as the Euclidean distance threshold.

[0081] Furthermore, the execution steps of the fitting prediction module 14 include: Extract the first-level cultivation sample set and the second-level cultivation sample set of the first trait index attribute; Group the second-level cultivation sample set according to the first-level cultivation sample set to obtain multiple groups of second-level cultivation samples; Traverse the multiple groups of second-level cultivation samples to calculate the central values of the trait indexes respectively, and obtain multiple second-level trait index central values; After calculating the mean value of the trait index between the multiple second-level trait index central values and the corresponding first-level cultivation sample set, calculate the central value of the trait index again to obtain the predicted value of the first cultivation trait index, and add it to the first group of predicted values of the cultivation trait index, and add it to the predicted values of the several groups of cultivation trait indexes.

[0082] Furthermore, the execution steps of the optimization module 15 include: Configure the coordinate search rule: According to the several trait deviation coefficients from small to large, sequentially extract the first medium distribution coordinate, the second medium distribution coordinate, and the third medium distribution coordinate from the several medium distribution coordinates; Calculate the coordinate mean values of the first medium distribution coordinate, the second medium distribution coordinate, and the third medium distribution coordinate to obtain the search reference coordinate; Based on the search reference coordinate, perform coordinate search on the several medium distribution coordinates to obtain the updated medium distribution coordinates and perform optimization iteration.

[0083] Embodiment 3. The embodiment of the present invention further provides an artificial intelligence umbilical cord stem cell cultivation parameter control device, including an artificial intelligence umbilical cord stem cell cultivation parameter control system.

[0084] It should be noted that in the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not detailedly described in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0085] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0086] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be realized by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for realizing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0087] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means realizes the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0088] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for realizing the functions specified in Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0089] Although the preferred embodiments of the present invention have been described, those skilled in the art can make additional changes and modifications to these embodiments once they know the basic inventive concept.

[0090] Obviously, those skilled in the art can make various changes and variations to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the present invention and its equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. An artificial intelligence umbilical cord stem cell cultivation parameter control method, characterized in that: include: Assigning values ​​to the culture intervention medium of umbilical cord stem cells to obtain several medium distribution coordinates; Traversing the plurality of medium distribution coordinates and combining experiments and literature to collect adjacent samples of trait indicators, and obtaining a primary cultivation sample set, wherein the primary cultivation sample includes a primary medium distribution coordinate set; Traversing the primary medium distribution coordinate set and combining experiments and literature to collect adjacent samples of trait indicators to obtain a secondary cultivation sample set; Based on the trait indices of the primary cultivation sample set and the secondary cultivation sample set, performing concentrated value fitting on the plurality of medium distribution coordinates, obtaining a plurality of groups of cultivation trait index prediction values, and calculating a plurality of individual trait deviation coefficients from the cultivation trait index expected values; The plurality of medium distribution coordinates are optimized according to the plurality of trait deviation coefficients, and target medium distribution coordinates are obtained and sent to the umbilical cord stem cell cultivation database.

2. The method according to claim 1, characterized in that Assign values ​​to the culture intervention medium of umbilical cord stem cells and obtain several medium distribution coordinates, including: The cultivation intervention medium includes discrete intervention medium and continuous intervention medium; Performing random assignment based on a discrete intervention medium constraint value set to obtain a first assignment result; Performing random assignment based on the continuous intervention medium constraint interval to obtain a second assignment result; The first assignment result and the second assignment result are combined to obtain a first medium distribution coordinate, which is added to the plurality of medium distribution coordinates.

3. The method according to claim 2, characterized in that Combining the first assignment result and the second assignment result to obtain first medium distribution coordinates includes: Obtain medium distribution coordinate system and abnormal detection materials; Combining the first assignment result and the second assignment result, placing them into the medium distribution coordinate system, and obtaining initial medium distribution coordinates; Performing anomaly detection on the initial medium distribution coordinates by using the anomaly detection material to obtain a coordinate anomaly degree; When the coordinate anomaly degree is equal to 0, the initial medium distribution coordinates are set as the first medium distribution coordinates.

4. The method according to claim 3, characterized in that Performing anomaly detection on the initial medium distribution coordinates by using the anomaly detection material to obtain the coordinate anomaly degree includes: Step 1: The abnormal detection material includes a plurality of abnormal distribution coordinates, wherein any one of the plurality of abnormal distribution coordinates includes at least two groups, and the plurality of abnormal distribution coordinates are obtained by cultivation abnormality parameters collected through integrated networking; Step 2: Add the initial medium distribution coordinates to the several abnormal distribution coordinates to obtain an abnormal detection distribution coordinate set; Step 3: performing a plane cutting and bisection on the anomaly detection distribution coordinate set to obtain a single-sided anomaly detection distribution coordinate set having the initial medium distribution coordinates, which is set as a primary anomaly detection distribution coordinate set; Step 4: When the number of the first-level anomaly detection distribution coordinate sets is not equal to 1, and the number of binary divisions is less than or equal to a preset number, return to step 3 to execute a loop based on the first-level anomaly detection distribution coordinate sets; Step 5: When the number of the first-level anomaly detection distribution coordinate sets is equal to 1, the coordinate anomaly degree is configured to 0; when the number of the first-level anomaly detection distribution coordinate sets is not equal to 1 and the number of binary divisions is greater than a preset number, the coordinate anomaly degree is configured to 1.

5. The method according to claim 1, characterized in that Traversing the several medium distribution coordinates and combining experiments and literature to collect adjacent samples of trait indicators, a primary cultivation sample set is obtained, including: Obtain the attribute of the trait index to be analyzed; Based on the trait index attributes to be analyzed, a cultivation weight distribution is performed on the intervention medium attribute set through grey correlation analysis to obtain an intervention medium weight distribution; Based on the weight distribution of the intervention medium and combined with the Euclidean distance threshold, the plurality of medium distribution coordinates are traversed to construct a plurality of adjacent constraint distances, and the first-level first cultivation sample set belonging to the trait index to be analyzed is collected in combination with experiments and literature, and added to the first-level cultivation sample set.

6. The method according to claim 5, characterized in that Based on the weight distribution of the intervention medium, combined with the Euclidean distance threshold, traverse the plurality of medium distribution coordinates to construct a plurality of adjacent constraint distances, and collect the first-level first cultivation sample set belonging to the trait index to be analyzed in combination with experiments and literature, which also includes: Based on the intervention medium weight distribution, a plurality of historical intervention medium Euclidean distances when the attribute fluctuation value of the trait index to be analyzed is less than or equal to the fluctuation threshold value of the trait index to be analyzed are counted, wherein any one of the plurality of historical intervention medium Euclidean distances has a trigger frequency identifier; The minimum Euclidean distance value whose trigger frequency identifier is greater than or equal to the trigger frequency threshold is extracted from the plurality of historical intervention medium Euclidean distances and is set as the Euclidean distance threshold.

7. The method according to claim 1, characterized in that Based on the trait indicators of the primary cultivation sample set and the secondary cultivation sample set, performing concentrated value fitting on the plurality of medium distribution coordinates to obtain a plurality of groups of cultivation trait indicator prediction values, including: Extracting the primary cultivation sample set and the secondary cultivation sample set of the first trait index attribute; Grouping the secondary cultivation sample set according to the primary cultivation sample set to obtain multiple groups of secondary cultivation samples; Traversing the plurality of groups of secondary cultivation samples and calculating the concentrated values ​​of the trait indicators respectively, to obtain a plurality of concentrated values ​​of the secondary trait indicators; After calculating the mean of the trait indicators of the multiple secondary trait indicator concentration values ​​and the one-to-one corresponding primary cultivation sample set, the trait indicator concentration value is calculated again to obtain the first cultivation trait indicator prediction value, add it into the first group of cultivation trait indicator prediction values, and add it into the several groups of cultivation trait indicator prediction values.

8. The method according to claim 1, characterized in that Optimizing the plurality of medium distribution coordinates according to the plurality of trait deviation coefficients, obtaining target medium distribution coordinates and sending them to the umbilical cord stem cell cultivation database, including: Configure coordinate search rules: According to the plurality of characteristic deviation coefficients from small to large, sequentially extracting a first medium distribution coordinate, a second medium distribution coordinate and a third medium distribution coordinate from the plurality of medium distribution coordinates; Calculating the coordinate mean of the first medium distribution coordinate, the second medium distribution coordinate and the third medium distribution coordinate to obtain a search reference coordinate; A coordinate search is performed on the plurality of medium distribution coordinates based on the search reference coordinates to obtain updated medium distribution coordinates and perform optimization iteration.

9. An artificial intelligence umbilical cord stem cell cultivation parameter control system, characterized in that: For implementing the method according to any one of claims 1 to 8, the system comprises: A medium assignment module is used to assign values ​​to the culture intervention medium of umbilical cord stem cells to obtain a number of medium distribution coordinates; A sample collection module is used to traverse the plurality of medium distribution coordinates and collect adjacent samples of trait indicators in combination with experiments and literature to obtain a primary cultivation sample set, wherein the primary cultivation sample includes a primary medium distribution coordinate set; An adjacency collection module is used to traverse the primary medium distribution coordinate set and collect adjacent samples of trait indicators in combination with experiments and literature to obtain a secondary cultivation sample set; A fitting prediction module is used to perform concentrated value fitting on the plurality of medium distribution coordinates based on the trait indicators of the primary cultivation sample set and the secondary cultivation sample set, obtain a plurality of groups of cultivation trait indicator prediction values, and calculate a plurality of individual trait deviation coefficients from the cultivation trait indicator expected values; The optimization module is used to optimize the plurality of medium distribution coordinates according to the plurality of trait deviation coefficients, obtain target medium distribution coordinates and send them to the umbilical cord stem cell cultivation database.

10. An artificial intelligence umbilical cord stem cell cultivation parameter control device, characterized in that: An artificial intelligence umbilical cord stem cell cultivation parameter control system comprising the above-described embodiment.

Images