{"id":17487,"date":"2026-09-21T09:46:58","date_gmt":"2026-09-21T09:46:58","guid":{"rendered":"https:\/\/www.examlabs.com\/certification\/?p=17487"},"modified":"2026-09-21T09:46:58","modified_gmt":"2026-09-21T09:46:58","slug":"microsoft-ai-300-practice-test-questions-and-exam-dumps-part-6-q101-120","status":"publish","type":"post","link":"https:\/\/www.examlabs.com\/certification\/microsoft-ai-300-practice-test-questions-and-exam-dumps-part-6-q101-120\/","title":{"rendered":"Microsoft AI-300 Practice Test Questions and Exam Dumps Part 6 Q101-120"},"content":{"rendered":"

View Full <\/b>Microsoft AI-300 Exam Dumps<\/b><\/a> and Practice Test Dumps<\/b><\/p>\n

 <\/p>\n

Question 101. Which practice helps ensure that an AI\/ML deployment can be reproduced consistently across environments?<\/b><\/p>\n

    \n
  1. <\/b> Manually configuring each environment<\/span>
    \n<\/span>2.<\/b> Using version-controlled infrastructure and environment definitions<\/span>
    \n<\/span>3.<\/b> Installing dependencies only after deployment<\/span>
    \n<\/span>4.<\/b> Changing configurations directly in production<\/span><\/li>\n<\/ol>\n

    Correct Answer: 2. Using version-controlled infrastructure and environment definitions<\/b><\/p>\n

    Explanation:<\/b><\/p>\n

    Version-controlled infrastructure and environment definitions help ensure that deployments can be reproduced consistently across development, testing, and production environments. Instead of relying on undocumented manual configuration, the required infrastructure settings, dependencies, runtime versions, and deployment parameters are maintained as controlled artifacts. When a deployment needs to be recreated, the same definitions can be applied again. This reduces configuration drift and makes troubleshooting easier because teams can identify exactly what changed between versions. For AI\/ML systems, reproducibility is especially important because model behavior can depend on libraries, runtime versions, compute configuration, and other environmental factors.<\/span><\/p>\n

    Question 102. What is the primary purpose of implementing automated validation before promoting an ML model to production?<\/b><\/p>\n

      \n
    1. <\/b> To eliminate the need for monitoring<\/span>
      \n<\/span>2.<\/b> To increase manual configuration<\/span>
      \n<\/span>3.<\/b> To verify that the model meets predefined quality and security requirements<\/span>
      \n<\/span>4.<\/b> To permanently disable previous model versions<\/span><\/li>\n<\/ol>\n

      Correct Answer: 3. To verify that the model meets predefined quality and security requirements<\/b><\/p>\n

      Explanation:<\/b><\/p>\n

      Automated validation provides a consistent checkpoint before a model is promoted to a production environment. Validation can evaluate model performance, data quality, security requirements, schema compatibility, resource requirements, and other predefined criteria. If the model does not satisfy the required thresholds, the deployment pipeline can stop or pause rather than automatically releasing the model. This approach reduces the possibility of human error and ensures that every candidate model goes through the same checks. Automated validation is particularly useful in continuous delivery because frequent model updates can otherwise make manual review difficult to perform consistently.<\/span><\/p>\n

      Question 103. Which deployment strategy keeps an existing production version available while a new version is deployed separately?<\/b><\/p>\n

        \n
      1. <\/b> Blue-green deployment<\/span>
        \n<\/span>2.<\/b> Random deployment<\/span>
        \n<\/span>3.<\/b> Manual-only deployment<\/span>
        \n<\/span>4.<\/b> Single-stage deployment<\/span><\/li>\n<\/ol>\n

        Correct Answer: 1. Blue-green deployment<\/b><\/p>\n

        Explanation:<\/b><\/p>\n

        Blue-green deployment maintains two production-capable environments or versions. The blue environment represents the currently active version, while the green environment contains the new version being prepared for release. The new model can be deployed and validated in the inactive environment before production traffic is switched to it. If validation succeeds, traffic can be redirected to the new version. If a serious issue appears after the switch, traffic can potentially be redirected back to the previous environment. This strategy reduces deployment risk because the known-good version remains available while the replacement version is tested and prepared.<\/span><\/p>\n

        Question 104. Which metric is most useful for detecting whether an online AI endpoint is becoming slow for users?<\/b><\/p>\n

          \n
        1. <\/b> Model file size<\/span>
          \n<\/span>2.<\/b> Training dataset name<\/span>
          \n<\/span>3.<\/b> Number of source-control commits<\/span>
          \n<\/span>4.<\/b> Request latency<\/span><\/li>\n<\/ol>\n

          Correct Answer: 4. Request latency<\/b><\/p>\n

          Explanation:<\/b><\/p>\n

          Request latency measures how long an online endpoint takes to process and respond to a request. Monitoring latency is important because an AI service can remain operational while still providing poor user experience if responses become increasingly slow. Teams can establish baseline latency levels and configure monitoring thresholds to detect unusual increases. High latency may indicate insufficient compute capacity, increased traffic, inefficient model execution, dependency problems, or infrastructure issues. Monitoring request latency together with error rates, throughput, and resource utilization provides a more complete picture of endpoint health and helps teams identify performance degradation before it becomes a larger production problem.<\/span><\/p>\n

          Question 105. Why should model artifacts be associated with version information in an MLOps workflow?<\/b><\/p>\n

            \n
          1. <\/b> To make all models identical<\/span>
            \n<\/span>2.<\/b> To remove the need for testing<\/span>
            \n<\/span>3.<\/b> To identify exactly which model version was trained, evaluated, and deployed<\/span>
            \n<\/span>4.<\/b> To prevent models from being monitored<\/span><\/li>\n<\/ol>\n

            Correct Answer: 3. To identify exactly which model version was trained, evaluated, and deployed<\/b><\/p>\n

            Explanation:<\/b><\/p>\n

            Version information provides traceability throughout the machine learning lifecycle. When a model artifact is versioned, teams can determine which specific model was trained, evaluated, approved, and ultimately deployed. This becomes especially important when several model iterations exist and production behavior needs to be investigated. Versioning can also support rollback because a previously validated model can be identified and redeployed when necessary. Without reliable version information, teams may struggle to determine which artifact generated particular predictions or whether a production issue was caused by the model itself or another component of the deployment.<\/span><\/p>\n

            Question 106. Which approach best protects secrets required by an AI deployment pipeline?<\/b><\/p>\n

              \n
            1. <\/b> Store passwords directly in source code<\/span>
              \n<\/span>2.<\/b> Put secrets in public configuration files<\/span>
              \n<\/span>3.<\/b> Share credentials through deployment logs<\/span>
              \n<\/span>4.<\/b> Use a managed secret store with appropriate identity-based access<\/span><\/li>\n<\/ol>\n

              Correct Answer: 4. Use a managed secret store with appropriate identity-based access<\/b><\/p>\n

              Explanation:<\/b><\/p>\n

              A managed secret store provides a safer mechanism for handling credentials, keys, certificates, and other sensitive configuration values. Instead of embedding secrets in application code or configuration files, the deployment can retrieve them when required using an authorized identity. Access permissions should follow the principle of least privilege so that services receive only the permissions necessary for their tasks. Centralized secret management also makes credential rotation and auditing easier. For production AI workloads, avoiding hard-coded credentials reduces the risk that sensitive information will accidentally be exposed through source repositories, deployment artifacts, logs, or other development and operational systems.<\/span><\/p>\n

              Question 107. What should happen when an automated model quality gate fails during deployment?<\/b><\/p>\n

                \n
              1. <\/b> The model should automatically replace the production model<\/span>
                \n<\/span>2.<\/b> The pipeline should stop or pause promotion until the issue is reviewed<\/span>
                \n<\/span>3.<\/b> All monitoring should be disabled<\/span>
                \n<\/span>4.<\/b> The failed test should be deleted<\/span><\/li>\n<\/ol>\n

                Correct Answer: 2. The pipeline should stop or pause promotion until the issue is reviewed<\/b><\/p>\n

                Explanation:<\/b><\/p>\n

                A quality gate is designed to prevent an unacceptable model from progressing automatically into production. If a model fails a required threshold, such as accuracy, precision, recall, data-quality validation, security checks, or another defined criterion, the pipeline should stop or pause promotion. This creates an opportunity for the team to investigate why the model failed and determine whether the problem can be corrected. Automatically ignoring a failed gate defeats its purpose and increases deployment risk. Quality gates are therefore an important control in automated MLOps pipelines because they combine repeatable testing with controlled promotion decisions.<\/span><\/p>\n

                Question 108. Which capability helps an organization compare a newly deployed model against the existing production model using real traffic?<\/b><\/p>\n

                  \n
                1. <\/b> Traffic splitting<\/span>
                  \n<\/span>2.<\/b> Source-code formatting<\/span>
                  \n<\/span>3.<\/b> Database indexing<\/span>
                  \n<\/span>4.<\/b> Manual documentation<\/span><\/li>\n<\/ol>\n

                  Correct Answer: 1. Traffic splitting<\/b><\/p>\n

                  Explanation:<\/b><\/p>\n

                  Traffic splitting allows production requests to be distributed between multiple deployed model versions according to defined percentages or routing rules. For example, a small portion of traffic can be directed to a new candidate model while the established version continues serving most requests. Teams can then compare performance, latency, error rates, prediction behavior, and other operational metrics. This controlled approach provides production evidence without immediately moving all users to the new model. Traffic splitting can support canary-style releases and controlled experimentation while maintaining a path to return traffic to the established version if unexpected behavior is detected.<\/span><\/p>\n

                  Question 109. Why should deployment configurations be stored under source control?<\/b><\/p>\n

                    \n
                  1. <\/b> To prevent configuration changes from being recorded<\/span>
                    \n<\/span>2.<\/b> To allow uncontrolled production changes<\/span>
                    \n<\/span>3.<\/b> To provide version history and traceability for deployment settings<\/span>
                    \n<\/span>4.<\/b> To eliminate deployment automation<\/span><\/li>\n<\/ol>\n

                    Correct Answer: 3. To provide version history and traceability for deployment settings<\/b><\/p>\n

                    Explanation:<\/b><\/p>\n

                    Source control provides a historical record of deployment configuration changes. When settings such as infrastructure definitions, endpoint configuration, pipeline parameters, or environment specifications are versioned, teams can identify who changed them, when the change occurred, and what the previous configuration looked like. This supports troubleshooting and improves reproducibility. If a deployment begins behaving differently after a configuration update, engineers can compare versions and identify the relevant modification. Source control also enables peer review and automated deployment workflows, reducing the need for undocumented manual changes directly in production environments.<\/span><\/p>\n

                    Question 110. Which monitoring signal can help identify that incoming production data has changed from the data used during model development?<\/b><\/p>\n

                      \n
                    1. <\/b> Data drift<\/span>
                      \n<\/span>2.<\/b> CPU brand<\/span>
                      \n<\/span>3.<\/b> Source-code indentation<\/span>
                      \n<\/span>4.<\/b> Deployment filename<\/span><\/li>\n<\/ol>\n

                      Correct Answer: 1. Data drift<\/b><\/p>\n

                      Explanation:<\/b><\/p>\n

                      Data drift occurs when the statistical characteristics of incoming production data change compared with the reference data used during model development or validation. Monitoring for drift can reveal changes in feature distributions, categories, ranges, or other characteristics that may affect model behavior. A model that performed well on historical data may become less reliable when real-world inputs change significantly. Drift monitoring does not automatically prove that model performance has degraded, but it provides an important warning signal that further investigation may be required. Teams can combine drift metrics with model-performance and data-quality measurements for a more complete monitoring strategy.<\/span><\/p>\n

                      Question 111. What is the main purpose of maintaining deployment logs for an AI production service?<\/b><\/p>\n

                        \n
                      1. <\/b> To replace source control<\/span>
                        \n<\/span>2.<\/b> To provide operational evidence for troubleshooting and auditing<\/span>
                        \n<\/span>3.<\/b> To permanently store user passwords<\/span>
                        \n<\/span>4.<\/b> To prevent future deployments<\/span><\/li>\n<\/ol>\n

                        Correct Answer: 2. To provide operational evidence for troubleshooting and auditing<\/b><\/p>\n

                        Explanation:<\/b><\/p>\n

                        Deployment logs provide useful operational evidence about what occurred during a release. They can record information such as deployment stages, versions, configuration changes, validation results, errors, warnings, and timing information. When a production deployment fails or behaves unexpectedly, these records help engineers understand where the process encountered a problem. Logs can also support auditing by showing important deployment events and actions. They should be designed carefully so that sensitive credentials, secrets, or unnecessary personal information are not exposed. Effective logging therefore improves both operational troubleshooting and accountability across the AI deployment lifecycle.<\/span><\/p>\n

                        Question 112. Which practice most directly supports rapid recovery after a failed production model deployment?<\/b><\/p>\n

                          \n
                        1. <\/b> Removing previous model versions<\/span>
                          \n<\/span>2.<\/b> Disabling monitoring<\/span>
                          \n<\/span>3.<\/b> Keeping a known-good model version and rollback procedure<\/span>
                          \n<\/span>4.<\/b> Changing several unrelated configurations simultaneously<\/span><\/li>\n<\/ol>\n

                          Correct Answer: 3. Keeping a known-good model version and rollback procedure<\/b><\/p>\n

                          Explanation:<\/b><\/p>\n

                          A known-good model version combined with a tested rollback procedure allows an organization to restore a stable production state when a deployment causes problems. The rollback process may involve redirecting traffic to a previous model, restoring an earlier endpoint configuration, or redeploying a validated artifact. Maintaining versioned models and deployment configurations makes this process more reliable because the recovery target is clearly identified. Automated rollback can further reduce recovery time when predefined failure conditions are detected. The goal is not simply to deploy quickly, but to ensure that failures can be contained and production service can be restored in a controlled manner.<\/span><\/p>\n

                          Question 113. Which testing type verifies that multiple components of an AI application work correctly together?<\/b><\/p>\n

                            \n
                          1. <\/b> Unit testing<\/span>
                            \n<\/span>2.<\/b> Integration testing<\/span>
                            \n<\/span>3.<\/b> Formatting validation<\/span>
                            \n<\/span>4.<\/b> Naming validation<\/span><\/li>\n<\/ol>\n

                            Correct Answer: 2. Integration testing<\/b><\/p>\n

                            Explanation:<\/b><\/p>\n

                            Integration testing verifies that different components of a system interact correctly. In an AI application, this might include interactions between an inference service, model artifact, data-processing component, authentication mechanism, storage service, or monitoring system. A component may pass its individual unit tests while still failing when connected to another service because of interface mismatches, incorrect schemas, authentication problems, or incompatible assumptions. Integration tests help identify these issues before production deployment. They are therefore an important part of an automated delivery pipeline, particularly for AI systems that depend on several interconnected services and infrastructure components.<\/span><\/p>\n

                            Question 114. What is the purpose of an ML pipeline component that can be reused across multiple workflows?<\/b><\/p>\n

                              \n
                            1. <\/b> To increase duplicated implementation<\/span>
                              \n<\/span>2.<\/b> To eliminate version control<\/span>
                              \n<\/span>3.<\/b> To provide a standardized, repeatable unit of processing<\/span>
                              \n<\/span>4.<\/b> To prevent automation<\/span><\/li>\n<\/ol>\n

                              Correct Answer: 3. To provide a standardized, repeatable unit of processing<\/b><\/p>\n

                              Explanation:<\/b><\/p>\n

                              Reusable pipeline components allow common machine learning operations to be defined once and used across multiple workflows. A component might perform data preparation, validation, feature processing, model evaluation, or another repeatable task. Reuse reduces duplication and helps teams maintain consistent behavior across different pipelines. Components can also be versioned so that changes are traceable and previous implementations remain available when required. This approach supports maintainability and reproducibility while making complex workflows easier to construct. Instead of rebuilding the same processing logic for every experiment or deployment, teams can compose standardized components into repeatable automated pipelines.<\/span><\/p>\n

                              Question 115. Why should production model performance be monitored after deployment?<\/b><\/p>\n

                                \n
                              1. <\/b> Because validation results from development never matter<\/span>
                                \n<\/span>2.<\/b> Because real-world data and behavior can change over time<\/span>
                                \n<\/span>3.<\/b> Because production models cannot be versioned<\/span>
                                \n<\/span>4.<\/b> Because monitoring replaces testing<\/span><\/li>\n<\/ol>\n

                                Correct Answer: 2. Because real-world data and behavior can change over time<\/b><\/p>\n

                                Explanation:<\/b><\/p>\n

                                A model that performs well during development may behave differently after deployment because real-world inputs, user behavior, data distributions, and operating conditions can change. Production monitoring helps detect performance degradation, unusual prediction patterns, data drift, latency increases, error rates, and other operational issues. Monitoring allows teams to compare actual behavior with established baselines and thresholds. It also provides evidence for deciding when investigation, retraining, configuration changes, or rollback may be necessary. Continuous monitoring is therefore an important part of the MLOps lifecycle because deployment is not the end of the model lifecycle.<\/span><\/p>\n

                                Question 116. Which identity approach is generally preferred for allowing an Azure-hosted service to access another Azure resource securely without embedding credentials?<\/b><\/p>\n

                                  \n
                                1. <\/b> Managed identity<\/span>
                                  \n<\/span>2.<\/b> Hard-coded password<\/span>
                                  \n<\/span>3.<\/b> Shared administrator account<\/span>
                                  \n<\/span>4.<\/b> Credential stored in source code<\/span><\/li>\n<\/ol>\n

                                  Correct Answer: 1. Managed identity<\/b><\/p>\n

                                  Explanation:<\/b><\/p>\n

                                  Managed identities allow supported Azure resources and applications to authenticate to other Azure services without requiring developers to embed passwords, keys, or client secrets directly into application code. Access can then be controlled through Azure role assignments and permissions. This approach reduces credential-management overhead and lowers the risk associated with accidentally exposing long-lived secrets. Managed identities can be especially useful in automated AI pipelines that need access to storage, secret stores, model artifacts, or other cloud resources. Permissions should still follow least-privilege principles, because eliminating stored credentials does not eliminate the need for careful authorization.<\/span><\/p>\n

                                  Question 117. Which practice provides evidence that a deployed model was created from a specific code and configuration state?<\/b><\/p>\n

                                    \n
                                  1. <\/b> Deleting pipeline history<\/span>
                                    \n<\/span>2.<\/b> Versioning code, configuration, model artifacts, and relevant metadata<\/span>
                                    \n<\/span>3.<\/b> Editing production files without records<\/span>
                                    \n<\/span>4.<\/b> Using identical filenames for every release<\/span><\/li>\n<\/ol>\n

                                    Correct Answer: 2. Versioning code, configuration, model artifacts, and relevant metadata<\/b><\/p>\n

                                    Explanation:<\/b><\/p>\n

                                    Traceability requires teams to connect a deployed model with the code, configuration, dependencies, and other artifacts that produced it. Versioning these elements creates a record of the exact state associated with a particular release. This is valuable when investigating production issues, reproducing results, comparing model versions, or performing a rollback. Metadata can also capture information such as training runs, evaluation results, data references, and deployment details. Together, these records provide a stronger audit trail than relying on filenames or undocumented manual processes. Traceability is a core characteristic of a controlled and reproducible MLOps workflow.<\/span><\/p>\n

                                    Question 118. What should an automated deployment pipeline do when a newly deployed endpoint fails its health checks?<\/b><\/p>\n

                                      \n
                                    1. <\/b> Ignore the health-check failure<\/span>
                                      \n<\/span>2.<\/b> Increase traffic immediately<\/span>
                                      \n<\/span>3.<\/b> Treat the deployment as unhealthy and trigger the defined recovery or failure process<\/span>
                                      \n<\/span>4.<\/b> Delete all monitoring data<\/span><\/li>\n<\/ol>\n

                                      Correct Answer: 3. Treat the deployment as unhealthy and trigger the defined recovery or failure process<\/b><\/p>\n

                                      Explanation:<\/b><\/p>\n

                                      Health checks help determine whether a deployed endpoint is functioning sufficiently to receive production traffic. If health checks fail, the deployment pipeline should treat the new version as unhealthy rather than automatically considering the release successful. Depending on the deployment design, the pipeline may stop promotion, keep traffic on the previous version, retry initialization, or trigger an automated rollback. This protects users from being routed to an unhealthy service. Health checks are particularly valuable when combined with deployment gates because they provide an objective operational signal that can be evaluated automatically before and after traffic is shifted.<\/span><\/p>\n

                                      Question 119. Which approach helps reduce deployment risk by exposing a new model to only a small percentage of users initially?<\/b><\/p>\n

                                        \n
                                      1. <\/b> Canary deployment<\/span>
                                        \n<\/span>2.<\/b> Immediate full rollout<\/span>
                                        \n<\/span>3.<\/b> Manual configuration replacement<\/span>
                                        \n<\/span>4.<\/b> Unversioned deployment<\/span><\/li>\n<\/ol>\n

                                        Correct Answer: 1. Canary deployment<\/b><\/p>\n

                                        Explanation:<\/b><\/p>\n

                                        Canary deployment introduces a new model gradually by routing a limited portion of production traffic to it while the existing version continues serving most requests. Teams can monitor the new version for errors, latency, resource usage, prediction behavior, and other defined metrics before increasing its traffic share. If the candidate performs as expected, traffic can be increased in controlled stages. If problems appear, the organization can reduce or remove traffic from the new version and preserve the established model. This strategy reduces exposure to deployment failures and provides real-world operational evidence before a broader release.<\/span><\/p>\n

                                        Question 120. Which combination best represents a mature AI\/ML production lifecycle?<\/b><\/p>\n

                                          \n
                                        1. <\/b> Manual deployment, no monitoring, and permanent model replacement<\/span>
                                          \n<\/span>2.<\/b> Development-only testing followed by unrestricted production release<\/span>
                                          \n<\/span>3.<\/b> Version control, automated testing, controlled deployment, monitoring, and rollback<\/span>
                                          \n<\/span>4.<\/b> Production changes without documentation or validation<\/span><\/li>\n<\/ol>\n

                                          Correct Answer: 3. Version control, automated testing, controlled deployment, monitoring, and rollback<\/b><\/p>\n

                                          Explanation:<\/b><\/p>\n

                                          A mature AI\/ML production lifecycle combines engineering controls across the entire model lifecycle rather than focusing only on training. Version control provides traceability for code and configurations, while automated testing validates components before deployment. Controlled deployment methods such as canary or blue-green releases reduce exposure to production failures. Continuous monitoring provides visibility into endpoint health, latency, errors, data drift, and model behavior after release. Finally, a documented and preferably automated rollback process provides a recovery path when predefined failure conditions occur. Together, these practices create a repeatable and controlled workflow that supports reliable AI operations.<\/span><\/p>\n

                                           <\/p>\n","protected":false},"excerpt":{"rendered":"

                                          View Full Microsoft AI-300 Exam Dumps and Practice Test Dumps   Question 101. Which practice helps ensure that an AI\/ML deployment can be reproduced consistently across environments? Manually configuring each environment 2. Using version-controlled infrastructure and environment definitions 3. Installing dependencies only after deployment 4. Changing configurations directly in production Correct Answer: 2. Using version-controlled […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":[],"categories":[1648,1647],"tags":[],"_links":{"self":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/17487"}],"collection":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/comments?post=17487"}],"version-history":[{"count":1,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/17487\/revisions"}],"predecessor-version":[{"id":17488,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/posts\/17487\/revisions\/17488"}],"wp:attachment":[{"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/media?parent=17487"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/categories?post=17487"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.examlabs.com\/certification\/wp-json\/wp\/v2\/tags?post=17487"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}