--- title: "PUT /udl/aiprediction/{id}" method: PUT path: "/udl/aiprediction/{id}" tags: ["Models and Simulations"] --- # PUT /udl/aiprediction/{id} `PUT /udl/aiprediction/{id}` ## Request - Base URL: `https://unifieddatalibrary.com` - URL: `https://unifieddatalibrary.com/udl/aiprediction/{id}` - Auth: none declared Service operation to update a single AIPrediction record. A specific role is required to perform this service operation. Please contact the UDL team for assistance. ## Path parameters - `id` string, required ## Request body - AIPredictionIngest: Prediction results generated by an AI model. Describes the data and relationships involved, along with explanations and uncertainty estimates that provide transparency into how the prediction was made. - `id` string: Unique identifier of the record, auto-generated by the system. - `classificationMarking` string, required: Classification marking of the data in IC/CAPCO Portion-marked format. - `idAIModel` string, required: Unique identifier of the model associated with the prediction. - `predictionId` string, required: Unique identifier for this prediction. This may be an internal identifier and not necessarily map to a valid identifier within the UDL. - `predictionTime` string, date-time, required: Datetime this prediction was made, in ISO 8601 UTC with millisecond precision. - `predictionSource` string: Source or origin of this prediction. - `prePipelineVersion` string: Version of the preprocessing pipeline used prior to this prediction. The pipeline is the structured sequence of steps applied to the input data to convert it to a standardized format. - `rte` string[]: List of runtime environment components used for this prediction, such as hardware, operating system, etc. - `batchSize` integer: Number of input items processed together in a single inference batch. - `latency` number, double: Total time taken to complete the prediction request in milliseconds. - `aiPredictionReviews` AIPredictionReviewIngest[]: Collection of evaluations performed on this prediction. - `status` string: Review status of the prediction (e.g., APPROVED, IN REVIEW, PENDING, REJECTED). - `reviewer` string: Identifies the person or system that reviewed the prediction. - `reviewDate` string, date-time: Datetime when the prediction review was completed, in ISO 8601 UTC with millisecond precision. - `corrections` string[]: One or more labels or annotations provided by the reviewer that correct the model's output. - `notes` string: Notes associated with the prediction review. - `aiPredictionEntities` AIPredictionEntityIngest[]: Collection of entities involved in this prediction. An entity represents a physical or digital data source used for a prediction, such as a digital file, image, etc. - `entitySourceType` string: Type of source that produced the entity, such as a sensor, satellite, platform, etc. - `entitySourceId` string: Identifier of the entity source. - `entitySourceName` string: Name of the entity source. - `entitySourceDetails` string: Other details about the entity source. - `entityType` string: Type of entity involved in the prediction (e.g., IMAGE, TEXT, VIDEO, etc.). - `entityId` string: Identifier of the entity involved in the prediction. - `entityRole` string: The role of the entity involved in the prediction (e.g., INPUT, OUTPUT, REFERENCE, etc.). - `entityURI` string: URI location of the entity involved in this prediction. - `mimeType` string: Indicates the MIME (Multipurpose Internet Mail Extensions) type of the entity (e.g., IMAGE/PNG, TEXT/PLAIN, etc.). - `checksumAlgorithm` string: Name of the algorithm used to generate the checksum for this entity (e.g., SHA-256, SHA-512, etc.). - `checksumValue` string: The result of applying the specified checksumAlgorithm to the entity. - `aiPredictionAssociations` AIPredictionAssociationIngest[]: Collection of model prediction associations related to this prediction. An association gives context and meaning to a prediction. - `assocEntityId` string: Identifier of the entity involved in this association, if applicable. - `assocData` string: Specifies what data is associated to the prediction (e.g., DOCUMENT, IMAGE, URN ADDRESS, etc.). - `assocType` string: Specifies what type of data is associated to the prediction (e.g., EXPLANATION, INPUT, REFERENCE, etc.). - `assocContext` string: Describes why the data is associated to the prediction. - `assocDataKeys` string[]: Array of keys that provide specific details about the data associated to the prediction, such as bounding box coordinates, time ranges, feature weights, etc. The entries in this array must correspond to the position index in the assocDataValues and assocDataUofMs arrays. This array must be the same length as assocDataValues and assocDataUofMs. - `assocDataValues` number[]: Array of values for the assocDataKeys that provide specific details about the data associated to the prediction, such as bounding box coordinates, time ranges, feature weights, etc. The entries in this array must correspond to the position index in the assocDataKeys and assocDataUofMs arrays. This array must be the same length as assocDataKeys and assocDataUofMs. - `assocDataUofMs` string[]: Array of units for the assocDataValues that provide specific details about the data associated to the prediction, such as bounding box coordinates, time ranges, feature weights, etc. The entries in this array must correspond to the position index in the assocDataKeys and assocDataValues arrays. This array must be the same length as assocDataKeys and assocDataValues. If there are no physical units associated with the assocDataValues, a value of NONE should be specified. - `assocTimestamp` string, date-time: Datetime the association between the data and the prediction was established, in ISO 8601 UTC with millisecond precision. - `aiPredictionContents` AIPredictionContentIngest[], required: Collection of model prediction content details related to this prediction. Content is the model-generated output based on a prediction. - `contentOutput` string, required: The content output generated from the prediction. - `contentType` string: Type of content generated from the prediction (e.g., BOUNDING BOX, LABEL, PROBABILITY, etc.). - `confidenceScore` number, double: Score between 0 and 1 that represents the model's certainty or probability that its output is correct. - `aiPredictionContentExplanations` AIPredictionContentExplanationIngest[]: Collection of model prediction explanations related to the prediction content. An explanation describes how or why a specific prediction content was produced by the AI model. - `explanationType` string: Category of explanation technique used in the prediction (e.g., ACTIVATION MAP, SALIENCY MAP, SHAP, etc.). - `explanationFormat` string: Format of the explanation (e.g., IMAGE, TABLE, TEXT, etc.). - `aiPredictionContentExplanationData` AIPredictionContentExplanationDataIngest[]: Collection of data that supports this prediction explanation. - `elementType` string: The type of element the explanation data describes (e.g., FEATURE, REGION, TOKEN, etc.). - `elementName` string: Identifies the element being explained (AGE, SIZE, VOLTAGE, etc.). - `elementValue` number, double: Value associated with the element, expressed in the specified unit of measure (elementUoM). - `elementUofM` string: The unit of measure associated with the element's value. If there are no physical units associated with the measurement, a value of NONE should be specified. - `impactScore` number, double: Measured impact the element had on the prediction. Positive values are associated with a higher prediction probability, while negative values are associated with a lower prediction probability. - `explanationDataNotes` string: Notes associated with the explanation data. - `aiPredictionContentUncertainties` AIPredictionContentUncertaintyIngest[]: Collection of model prediction uncertainties related to the prediction content. An uncertainty is a measure of how uncertain the AI model is about a specific prediction output. - `uncertCategory` string: Category of the source of uncertainty affecting the prediction, such as data noise or limited model knowledge (e.g., ALEATORIC, EPISTEMIC, MIXED). - `uncertType` string: The statistical or mathematical representation of the uncertainty for the prediction (e.g., CONFIDENCE INTERVAL, STANDARD DEVIATION, VARIANCE, etc.). - `uncertDescription` string: Describes the type of uncertainty being reported. - `uncertMethod` string: Describes the method or process used to calculate the uncertainty values (e.g., BOOTSTRAP, ENSEMBLE, MC DROPOUT, etc.). - `confidenceLevel` number, double: Probability the true value is within the given lowerBound and upperBound, expressed as a percent. The range of values is 0 to 100, with 100 being highest level of confidence. - `lowerBound` number, double: The smallest value in the confidence interval for the prediction. - `upperBound` number, double: The largest value in the confidence interval for the prediction. - `ensembleMethod` string: The approach used if multiple models are combined (e.g., BAGGING, STACKING, etc.). - `ensembleSize` integer: Number of models in the ensemble. - `samplingMethod` string: The method used to collect samples (e.g., BOOTSTRAP, MONTE CARLO, etc.). - `samplingCount` integer: Number of predictions or samples used to calculate the uncertainty values. - `seeds` integer[]: List of random number starting values used. - `calibrationTechnique` string: Method used to adjust prediction probabilities (e.g., ISOTONIC, NONE, PLATT, TEMPERATURE SCALING, etc.). - `calibrationParameter` string: The main numeric value used by the calibration technique. - `ece` number, double: Expected Calibration Error score. Measures the average difference between predicted probabilities and actual outcomes. Lower values indicate more accurate and calibrated predictions. - `brierScore` number, double: Brier score for probability accuracy. Calculates the mean squared difference between predicted probabilities and actual outcomes. Lower scores indicate more accurate and calibrated predictions. - `nll` number, double: Negative Log-Likelihood score. Measures how unlikely the observed data is under the predicted probability distribution. Lower values indicate better agreement between predictions and actual outcomes. - `uncertKeys` string[]: Array of keys that provide specific details about uncertainty elements associated to the prediction content. The entries in this array must correspond to the position index in the uncertValues and uncertUofMs arrays. This array must be the same length as uncertValues and uncertUofMs. - `uncertValues` number[]: Array of values for the uncertKeys that provide specific details about uncertainty elements associated to the prediction content. The entries in this array must correspond to the position index in the uncertKeys and uncertUofMs arrays. This array must be the same length as uncertKeys and uncertUofMs. - `uncertUofMs` string[]: Array of units for the uncertValues that provide specific details about uncertainty elements associated to the prediction content. The entries in this array must correspond to the position index in the uncertKeys and uncertValues arrays. This array must be the same length as uncertKeys and uncertValues. If there are no physical units associated with the uncertValues, a value of NONE should be specified. - `uncertNotes` string: Notes associated with this uncertainty data. - `tags` string[]: Optional array of provider/source specific tags for this data, where each element is no longer than 32 characters, used for implementing data owner conditional access controls to restrict access to the data. Should be left null by data providers unless conditional access controls are coordinated with the UDL team. - `transactionId` string: Optional identifier to track a commercial or marketplace transaction executed to produce this data. - `createdAt` string, date-time: Time the row was created in the database, auto-populated by the system. - `createdBy` string: Application user who created the row in the database, auto-populated by the system. - `updatedAt` string, date-time: Time the row was last updated in the database, auto-populated by the system. - `updatedBy` string: Application user who updated the row in the database, auto-populated by the system. - `sourceDL` string: The source data library from which this record was received. This could be a remote or tactical UDL or another data library. If null, the record should be assumed to have originated from the primary Enterprise UDL. - `source` string, required: Source of the data. - `origin` string: Originating system or organization which produced the data, if different from the source. The origin may be different than the source if the source was a mediating system which forwarded the data on behalf of the origin system. If null, the source may be assumed to be the origin. - `origNetwork` string: The originating source network on which this record was created, auto-populated by the system. - `dataMode` 'REAL' | 'TEST' | 'SIMULATED' | 'EXERCISE', required: Indicator of whether the data is REAL, TEST, EXERCISE, or SIMULATED data: REAL: Data collected or produced that pertains to real-world objects, events, and analysis. TEST: Specific datasets used to evaluate compliance with specifications and requirements, and for validating technical, functional, and performance characteristics. EXERCISE: Data pertaining to a government or military exercise. The data may include both real and simulated data. SIMULATED: Synthetic data generated by a model to mimic real-world datasets. ## Response `204` Success ## Other responses - `400`: Bad Request - `401`: Invalid/Missing User Credentials - `403`: User Is Not Authorized - `404`: Item Not Found - `415`: MediaType Unsupported - `429`: Rate of Requests Exceeds Allowed Limit - `500`: Internal Server Error ## Changes > 12 revisions in range; 1 could not be searched. - **2026-03-25** `173e4aff893f`: 1 info - endpoint added [Operation history](https://skmtc.dev/bluestaq/apis/unified-data-library-services-api/changes/udl/aiprediction/:id/put.md) --- [API](https://skmtc.dev/bluestaq/apis/unified-data-library-services-api.md) - [All operations](https://skmtc.dev/bluestaq/apis/unified-data-library-services-api/llms.txt) - [OpenAPI document](https://skmtc.dev/bluestaq/apis/unified-data-library-services-api/revisions/bd17bc53681e?raw)