Google Professional Machine Learning Engineer

• #81

  You're working on an application to assist users with meal planning, aiming to employ machine learning for extracting ingredients and kitchen cookware from a corpus of recipes saved as unstructured text files. What approach should you take?

  • AA. Set up a text dataset on Vertex AI tailored for entity extraction. Establish two entities named "ingredient" and "cookware" and label a minimum of 200 instances for each. Train an AutoML entity extraction model to identify occurrences of these entities. Assess model performance using a holdout dataset.
  • BB. Establish a multi-label text classification dataset on Vertex AI. Create a testing dataset and label each recipe with its corresponding ingredients and cookware. Train a multi-class classification model and evaluate its performance using a holdout dataset.
  • CC. Utilize the Entity Analysis feature of the Natural Language API to extract ingredients and cookware from each recipe. Evaluate the model's performance using a prelabeled dataset.
  • DD. Establish a text dataset on Vertex AI optimized for entity extraction. Generate entities corresponding to the different ingredients and cookware present. Train an AutoML entity extraction model to recognize these entities and evaluate its performance using a holdout dataset.
  उत्तर दिखाएँ
• #82

  As you keep an eye on your model training and observe the GPU utilization, you come to realize that you're using a native synchronous implementation. Moreover, your training data is divided into several files, and you're eager to minimize the execution time of your input pipeline. What steps should you take to address this situation?

  • AA. Increase the CPU load
  • BB. Add caching to the pipeline
  • CC. Increase the network bandwidth
  • DD. Add parallel interleave to the pipeline
  उत्तर दिखाएँ
• #83

  You started working on a classification problem with time series data and achieved an area under the receiver operating characteristic curve (AUC ROC) value of 99% for training data after just a few experiments. You haven't explored using any sophisticated algorithms or spent any time on hyperparameter tuning. What should your next step be to identify and fix the problem?

  • AA. Address the model overfitting by using a less complex algorithm.
  • BB. Address data leakage by applying nested cross-validation during model training.
  • CC. Address data leakage by removing features highly correlated with the target value.
  • DD. Address the model overfitting by tuning the hyperparameters to reduce the AUC ROC value.
  उत्तर दिखाएँ
• #84

  You are pre-training a large language model on Google Cloud, which involves custom TensorFlow operations in the training loop, a large batch size, and several weeks of training. You aim to configure a training architecture that minimizes both training time and compute costs. What should you do?

  • AA. Implement a TPU Pod slice with -accelerator-type=v4-l28 using tf.distribute.TPUStrategy.
  • BB. Implement 8 workers of a2-megagpu-16g machines using tf.distribute.MultiWorkerMirroredStrategy.
  • CC. Implement 16 workers of c2d-highcpu-32 machines using tf.distribute.MirroredStrategy.
  • DD. Implement 16 workers of a2-highgpu-8g machines using tf.distribute.MultiWorkerMirroredStrategy.
  उत्तर दिखाएँ
• #85

  The marketing team at your organization has expressed the need to send biweekly scheduled emails to customers who are anticipated to spend above a variable threshold. This marks the marketing team's first foray into machine learning (ML), and you've been assigned the responsibility of overseeing the implementation. To address this requirement, you initiated a new Google Cloud project and leveraged Vertex AI Workbench to craft a solution that involves model training and batch inference using an XGBoost model, utilizing transactional data stored in Cloud Storage. Your goal is to establish an end-to-end pipeline that seamlessly delivers predictions to the marketing team in a secure manner while also optimizing for cost-efficiency and minimizing the need for extensive code maintenance. What steps should you take to achieve this objective?

  • AA. Create a scheduled pipeline on Vertex AI Pipelines that accesses the data from Cloud Storage, uses Vertex AI to perform training and batch prediction, and outputs a file in a Cloud Storage bucket that contains a list of all customer emails and expected spending.
  • BB. Create a scheduled pipeline on Cloud Composer that accesses the data from Cloud Storage, copies the data to BigQuery, uses BigQuery ML to perform training and batch prediction, and outputs a table in BigQuery with customer emails and expected spending.
  • CC. Create a scheduled notebook on Vertex AI Workbench that accesses the data from Cloud Storage, performs training and batch prediction on the managed notebook instance, and outputs a file in a Cloud Storage bucket that contains a list of all customer emails and expected spending.
  • DD. Create a scheduled pipeline on Cloud Composer that accesses the data from Cloud Storage, uses Vertex AI to perform training and batch prediction, and sends an email to the marketing team’s Gmail group email with an attachment that contains an encrypted list of all customer emails and expected spending.
  उत्तर दिखाएँ
• #86

  You have recently trained a scikit-learn model that you plan to deploy on Vertex AI. This model will support both online and batch prediction. You need to preprocess input data for model inference. You want to package the model for deployment while minimizing additional code.

  • AA. 1. Upload your model to the Vertex AI Model Registry by using a prebuilt scikit-learn prediction container. 2.Deploy your model to Vertex AI Endpoints, and create a Vertex AI batch prediction job that uses the instanceConfig.instanceType setting to transform your input data.
  • BB. 1. Wrap your model in a custom prediction routine (CPR), and build a container image from the CPR local model. 2.Upload your scikit-learn model container to Vertex AI Model Registry. 3.Deploy your model to Vertex AI Endpoints, and create a Vertex AI batch prediction job.
  • CC. 1. Create a custom container for your scikit-learn model. 2.Define a custom serving function for your model. 3.Upload your model and custom container to Vertex AI Model Registry. 4.Deploy your model to Vertex AI Endpoints, and create a Vertex AI batch prediction job.
  • DD. 1. Create a custom container for your scikit-learn model. 2.Upload your model and custom container to Vertex AI Model Registry. 3.Deploy your model to Vertex AI Endpoints, and create a Vertex AI batch prediction job that uses the instanceConfig.instanceType setting to transform your input data.
  उत्तर दिखाएँ
• #87

  Your company operates an application that gathers news articles from various online sources and delivers them to users. You require a recommendation model that can propose articles to readers based on their current reading material, suggesting articles that are similar. Which approach should you employ for this task?

  • AA. Create a collaborative filtering system that recommends articles to a user based on the user’s past behavior.
  • BB. Encode all articles into vectors using word2vec, and build a model that returns articles based on vector similarity.
  • CC. Build a logistic regression model for each user that predicts whether an article should be recommended to a user.
  • DD. Manually label a few hundred articles, and then train an SVM classifier based on the manually classified articles that categorizes additional articles into their respective categories.
  उत्तर दिखाएँ
• #88

  Your organization's call center has tasked you with developing a model to analyze customer sentiments in each call. With over one million calls received daily, the data is stored in Cloud Storage. It's imperative that the data remains within the region of the call's origin and that no Personally Identifiable Information (PII) is stored or analyzed. Additionally, the data science team utilizes a third-party visualization tool that requires an SQL ANSI-2011 compliant interface. Given these requirements, what components should you select for data processing and analytics to effectively design the data pipeline?

  • AA. 1 = Google Cloud Dataflow, 2 = Google BigQuery
  • BB. 1 = Google Cloud Pub/Sub, 2 = Google Cloud Datastore
  • CC. 1 = Google Cloud Dataflow, 2 = Google Cloud SQL
  • DD. 1 = Google Cloud Functions, 2 = Google Cloud SQL
  उत्तर दिखाएँ
• #89

  You have the task of designing a recommendation system for a new video streaming platform. Your goal is to suggest the next video for users to watch. After receiving approval from an AI Ethics team, you're ready to commence development. Although your company's video catalog contains valuable metadata (e.g., content type, release date, country), you currently lack historical user event data. How should you go about constructing the recommendation system for the initial product version?

  • AA. Launch the product without machine learning. Present videos to users alphabetically, and start collecting user event data so you can develop a recommender model in the future.
  • BB. Launch the product without machine learning. Use simple heuristics based on content metadata to recommend similar videos to users, and start collecting user event data so you can develop a recommender model in the future.
  • CC. Launch the product with machine learning. Use a publicly available dataset such as MovieLens to train a model using the Recommendations AI, and then apply this trained model to your data.
  • DD. Launch the product with machine learning. Generate embeddings for each video by training an autoencoder on the content metadata using TensorFlow. Cluster content based on the similarity of these embeddings, and then recommend videos from the same cluster.
  उत्तर दिखाएँ
• #90

  You've developed a custom ML model using scikit-learn, and you've encountered longer-than-expected training times. To address this issue, you've chosen to migrate your model to Vertex AI Training and are now looking for initial steps to enhance the training efficiency. What should be your first approach?

  • AA. Train your model in a distributed mode using multiple Compute Engine VMs.
  • BB. Train your model using Vertex AI Training with CPUs.
  • CC. Migrate your model to TensorFlow, and train it using Vertex AI Training.
  • DD. Train your model using Vertex AI Training with GPUs.
  उत्तर दिखाएँ
• #91

  You are a member of a data science team at a bank, tasked with building an ML model for predicting loan default risk. Your dataset, consisting of hundreds of millions of cleaned records, is stored in a BigQuery table. Your objective is to create and evaluate multiple models using TensorFlow and Vertex AI while ensuring that the data ingestion process is efficient and scalable. To achieve this, what steps should you take to minimize bottlenecks during data ingestion?

  • AA. Use the BigQuery client library to load data into a dataframe, and use tf.data.Dataset.from_tensor_slices() to read it.
  • BB. Export data to CSV files in Cloud Storage, and use tf.data.TextLineDataset() to read them.
  • CC. Convert the data into TFRecords, and use tf.data.TFRecordDataset() to read them.
  • DD. Use TensorFlow I/O’s BigQuery Reader to directly read the data.
  उत्तर दिखाएँ
• #92

  As an employee of a public transportation company, your task is to develop a model that estimates delay times across various transportation routes. This model must provide real-time predictions to users via an app. Given the influence of seasonal changes and population growth on data relevance, the model requires monthly retraining. What is the optimal way to set up the end-to-end architecture of this predictive model, adhering to Google's recommended best practices?

  • AA. Configure Kubeflow Pipelines to schedule your multi-step workflow, from training to deploying the model.
  • BB. Utilize a model trained and deployed on BigQuery ML, and trigger retraining using the scheduled query feature in BigQuery.
  • CC. Write a Cloud Functions script that launches training and deployment jobs on AI Platform, triggered by Cloud Scheduler.
  • DD. Employ Cloud Composer to programmatically schedule a Dataflow job that executes the workflow, from training to deploying the model.
  उत्तर दिखाएँ
• #93

  You have recently deployed a scikit-learn model to a Vertex AI endpoint and are now in the process of testing it with live production traffic. While monitoring the endpoint, you've noticed that the number of requests per hour is twice as high as initially expected throughout the day. Your goal is to ensure that the endpoint can efficiently scale to meet increased demand in the future, thus preventing users from experiencing high latency. What actions should you take to address this situation?

  • AA. Deploy two models to the same endpoint, and distribute requests among them evenly
  • BB. Configure an appropriate minReplicaCount value based on expected baseline traffic
  • CC. Set the target utilization percentage in the autoscailngMetricSpecs configuration to a higher value
  • DD. Change the model’s machine type to one that utilizes GPUs
  उत्तर दिखाएँ
• #94

  You oversee a team of data scientists who utilize a cloud-based backend system to submit training jobs. Managing this system has become challenging, and you aim to switch to a managed service. The data scientists on your team work with various frameworks, such as Keras, PyTorch, Theano, Scikit-learn, and custom libraries. What would be your recommended course of action?

  • AA. Use the AI Platform custom containers feature to receive training jobs using any framework.
  • BB. Configure Kubeflow to run on Google Kubernetes Engine and receive training jobs through TF Job.
  • CC. Create a library of VM images on Compute Engine, and publish these images on a centralized repository.
  • DD. Set up Slurm workload manager to receive jobs that can be scheduled to run on your cloud infrastructure.
  उत्तर दिखाएँ
• #95

  You've trained an XGBoost model for deployment on Vertex AI for online prediction. As you upload your model to Vertex AI Model Registry, you need to configure the explanation method for serving online prediction requests with minimal latency. Additionally, you want to receive alerts when feature attributions of the model significantly change over time. What steps should you take?

  • AA. 1. Specify sampled Shapley as the explanation method with a path count of 5. 2. Deploy the model to Vertex AI Endpoints. Create a Model Monitoring job using prediction drift as the monitoring objective.
  • BB. 1. Specify Integrated Gradients as the explanation method with a path count of 5. 2. Deploy the model to Vertex AI Endpoints. Create a Model Monitoring job using prediction drift as the monitoring objective.
  • CC. 1. Specify sampled Shapley as the explanation method with a path count of 50. 2. Deploy the model to Vertex AI Endpoints. Create a Model Monitoring job using training-serving skew as the monitoring objective.
  • DD. 1. Specify Integrated Gradients as the explanation method with a path count of 50. 2. Deploy the model to Vertex AI Endpoints. Create a Model Monitoring job using training-serving skew as the monitoring objective.
  उत्तर दिखाएँ
• #96

  You've created a Vertex AI ML pipeline that involves preprocessing and training stages, and each of these stages operates within distinct custom Docker images. Within your organization, GitHub and GitHub Actions are employed for continuous integration and continuous deployment (CI/CD) to perform unit and integration tests. To automate the model retraining process, you seek a workflow that can be triggered manually and automatically whenever new code is merged into the main branch. Your goal is to streamline the workflow while maintaining flexibility. How should you set up and configure the CI/CD workflow to achieve this?

  • AA. Trigger a Cloud Build workflow to run tests, build custom Docker images, push the images to Artifact Registry, and launch the pipeline in Vertex AI Pipelines.
  • BB. Trigger GitHub Actions to run the tests, launch a job on Cloud Run to build custom Docker images, push the images to Artifact Registry, and launch the pipeline in Vertex AI Pipelines.
  • CC. Trigger GitHub Actions to run the tests, build custom Docker images, push the images to Artifact Registry, and launch the pipeline in Vertex AI Pipelines.
  • DD. Trigger GitHub Actions to run the tests, launch a Cloud Build workflow to build custom Docker images, push the images to Artifact Registry, and launch the pipeline in Vertex AI Pipelines.
  उत्तर दिखाएँ
• #97

  You are tasked with developing a strategy to efficiently organize jobs, models, and versions on AI Platform for your team of over 50 data scientists. Which strategy should you opt for to ensure a clean and scalable organization?

  • AA. Set up restrictive IAM permissions on the AI Platform notebooks so that only a single user or group can access a given instance.
  • BB. Separate each data scientist's work into a different project to ensure that the jobs, models, and versions created by each data scientist are accessible only to that user.
  • CC. Use labels to organize resources into descriptive categories. Apply a label to each created resource so that users can filter the results by label when viewing or monitoring the resources.
  • DD. Set up a BigQuery sink for Cloud Logging logs that is appropriately filtered to capture information about AI Platform resource usage. In BigQuery, create a SQL view that maps users to the resources they are using.
  उत्तर दिखाएँ
• #98

  You are employed by a small company that has implemented an ML model using autoscaling on Vertex AI to provide online predictions within a production setting. Currently, the model handles approximately 20 prediction requests per hour, with an average response time of one second. However, you've recently retrained the same model using fresh data and are now conducting a canary test by directing approximately 10% of the production traffic to this updated model. During this canary test, you've observed that prediction requests for the new model are taking anywhere from 30 to 180 seconds to finish. What step should you take to address this issue?

  • AA. Submit a request to raise your project quota to ensure that multiple prediction services can run concurrently.
  • BB. Turn off auto-scaling for the online prediction service of your new model. Use manual scaling with one node always available.
  • CC. Remove your new model from the production environment. Compare the new model and existing model codes to identify the cause of the performance bottleneck.
  • DD. Remove your new model from the production environment. For a short trial period, send all incoming prediction requests to BigQuery. Request batch predictions from your new model, and then use the Data Labeling Service to validate your model’s performance before promoting it to production.
  उत्तर दिखाएँ
• #99

  You are required to construct an image classification model utilizing an extensive dataset that is stored within a Cloud Storage bucket. How should you proceed with this task?

  • AA. Use Vertex AI Pipelines with the Kubeflow Pipelines SDK to create a pipeline that reads the images from Cloud Storage and trains the model.
  • BB. Use Vertex AI Pipelines with TensorFlow Extended (TFX) to create a pipeline that reads the images from Cloud Storage and trains the model.
  • CC. Import the labeled images as a managed dataset in Vertex AI and use AutoML to train the model.
  • DD. Convert the image dataset to a tabular format using Dataflow Load the data into BigQuery and use BigQuery ML to train the model.
  उत्तर दिखाएँ
• #100

  You are employed at a social media company and have a requirement to create a no-code image classification model for an iOS mobile application, specifically designed for identifying fashion accessories. Your labeled dataset is stored in Cloud Storage. In this context, you aim to configure a training workflow that not only minimizes cost but also provides predictions with the lowest possible latency. How should you proceed?

  • AA. Train the model using Vertex AI AutoML, and register the model in Vertex AI Model Registry. Configure your mobile application to send batch requests during prediction.
  • BB. Train the model using Vertex AI AutoML Edge, and export it as a Core ML model. Configure your mobile application to use the .mlmodel file directly.
  • CC. Train the model using Vertex AI AutoML Edge, and export the model as a TFLite model. Configure your mobile application to use the .tflite file directly.
  • DD. Train the model using Vertex AI AutoML, and expose the model as a Vertex AI endpoint. Configure your mobile application to invoke the endpoint during prediction.
  उत्तर दिखाएँ

अपना स्कोर ट्रैक करना, समयबद्ध मॉक परीक्षा देना और AI व्याख्या पाना चाहते हैं? निःशुल्क खाता बनाएँ