PMLE practice questions 1–10

1. Question 1 · 364 in bank

   You are building an ML model to detect anomalies in real-time sensor data. You will use Pub/Sub to handle incoming requests. You want to store the results for analytics and visualization. How should you configure the pipeline?

   • A1 = Dataflow, 2 = AI Platform, 3 = BigQuery
   • B1 = DataProc, 2 = AutoML, 3 = Cloud Bigtable
   • C1 = BigQuery, 2 = AutoML, 3 = Cloud Functions
   • D1 = BigQuery, 2 = AI Platform, 3 = Cloud Storage

   Show answer

   Answer: A

   Explanation

   B. 1 = DataProc, 2 = AutoML, 3 = Cloud Bigtable - wrong because bigtable is not for analytics and dataflow is preferred over dataproc. C. 1 = BigQuery, 2 = AutoML, 3 = Cloud Functions - wrong because cloud function is not a data store product D. 1 = BigQuery, 2 = AI Platform, 3 = Cloud Storage - wrong because cloud storage is not best way to store data for analytics

   Source: Adapted from community discussion — ashii007

   References

   • Google Cloud documentation
   • Google Cloud documentation
   • Google developer documentation
   • Google Cloud documentation
   • Google Cloud documentation

   Next question

2. Question 2 · 364 in bank

   Your team needs to build a model that predicts whether images contain a driver's license, passport, or credit card. The data engineering team already built the pipeline and generated a dataset composed of 10,000 images with driver's licenses, 1,000 images with passports, and 1,000 images with credit cards. You now have to train a model with the following label map: [`˜drivers_license', `˜passport', `˜credit_card']. Which loss function should you use?

   • ACategorical hinge
   • BBinary cross-entropy
   • CCategorical cross-entropy
   • DSparse categorical cross-entropy

   Show answer

   Answer: C

   Explanation

   Use sparse categorical crossentropy when your classes are mutually exclusive (e.g. when each sample belongs exactly to one class) and categorical crossentropy when one sample can have multiple classes or labels are soft probabilities (like [0.5, 0.3, 0.2]).

   Source: Adapted from community discussion — gcp2021go

   References

   • Google developer documentation

   Next question

3. Question 3 · 364 in bank

   You are an ML engineer at a manufacturing company. You need to build a model that identifies defects in products based on images of the product taken at the end of the assembly line. You want your model to preprocess the images with lower computation to quickly extract features of defects in products. Which approach should you use to build the model?

   • AReinforcement learning
   • BRecommender system
   • CRecurrent Neural Networks (RNN)
   • DConvolutional Neural Networks (CNN)

   Show answer

   Answer: D

   References

   • Google developer documentation

   Next question

4. Question 4 · 364 in bank

   You are developing an ML model intended to classify whether X-ray images indicate bone fracture risk. You have trained a ResNet architecture on Vertex AI using a TPU as an accelerator, however you are unsatisfied with the training time and memory usage. You want to quickly iterate your training code but make minimal changes to the code. You also want to minimize impact on the model’s accuracy. What should you do?

   • AReduce the number of layers in the model architecture.
   • BReduce the global batch size from 1024 to 256.
   • CReduce the dimensions of the images used in the model.
   • DConfigure your model to use bfloat16 instead of float32.

   Show answer

   Answer: D

   Explanation

   TPUs are optimized for operations with bfloat16 data types. By switching from float32 to bfloat16, you can benefit from the TPU's hardware acceleration capabilities, leading to faster computation and reduced memory usage without significant changes to your code. While bfloat16 offers a lower precision compared to float32, it maintains a similar dynamic range. This means that the reduction in numerical precision is unlikely to have a substantial impact on the accuracy of your model, especially in the context of image classification tasks like bone fracture risk assessment in X-rays. While reducing the batch size can decrease memory usage, it can also affect the model's convergence and accuracy. Additionally, TPUs are highly efficient with large batch sizes, so reducing the batch size might not fully leverage the TPU's capabilities.

   Source: Adapted from community discussion — tavva_prudhvi

   References

   • Google Cloud documentation

   Next question

5. Question 5 · 364 in bank

   You have successfully deployed to production a large and complex TensorFlow model trained on tabular data. You want to predict the lifetime value (LTV) field for each subscription stored in the BigQuery table named subscription. subscriptionPurchase in the project named my-fortune500-company-project. You have organized all your training code, from preprocessing data from the BigQuery table up to deploying the validated model to the Vertex AI endpoint, into a TensorFlow Extended (TFX) pipeline. You want to prevent prediction drift, i.e., a situation when a feature data distribution in production changes significantly over time. What should you do?

   • AImplement continuous retraining of the model daily using Vertex AI Pipelines.
   • BAdd a model monitoring job where 10% of incoming predictions are sampled 24 hours.
   • CAdd a model monitoring job where 90% of incoming predictions are sampled 24 hours.
   • DAdd a model monitoring job where 10% of incoming predictions are sampled every hour.

   Show answer

   Answer: B

   Explanation

   Continuous retraining (option A) is not necessarily the best solution for preventing prediction drift, as it can be time-consuming and expensive. Instead, monitoring the performance of the model in production is a better approach. Option B is a good choice because it samples a small percentage of incoming predictions and checks for any significant changes in the feature data distribution over a 24-hour period. This allows you to detect any drift and take appropriate action to address it before it affects the model's performance. Options C and D are less effective because they either sample too many or too few predictions and/or at too frequent intervals.

   Source: Adapted from community discussion — tavva_prudhvi

   References

   • Google Cloud documentation
   • Google Cloud documentation

   Next question

6. Question 6 · 364 in bank

   You recently developed a deep learning model using Keras, and now you are experimenting with different training strategies. First, you trained the model using a single GPU, but the training process was too slow. Next, you distributed the training across 4 GPUs using tf.distribute.MirroredStrategy (with no other changes), but you did not observe a decrease in training time. What should you do?

   • ADistribute the dataset with tf.distribute.Strategy.experimental_distribute_dataset
   • BCreate a custom training loop.
   • CUse a TPU with tf.distribute.TPUStrategy.
   • DIncrease the batch size.

   Show answer

   Answer: D

   Explanation

   D: it is best https://www.tensorflow.org/guide/distributed_training#use_tfdistributestrategy_with_keras_modelfit Each epoch will then train faster as you add more GPUs. Typically, you would want to increase your batch size as you add more accelerators, C is rule out because of GPU A and B , as reading on https://www.tensorflow.org/guide/distributed_training#use_tfdistributestrategy_with_custom_training_loops To use custom loop , we have call If you are writing a custom training loop, you will need to call a few more methods, see the guide: Start by creating a tf.data.Dataset normally. Use tf.distribute.Strategy.experimental_distribute_dataset to convert a tf.data.Dataset to something that produces "per-replica" values. If you want to https://www.tensorflow.org/api_docs/python/tf/distribute/Strategy

   Source: Adapted from community discussion — John_Pongthorn

   Next question

7. Question 7 · 364 in bank

   You work for a gaming company that has millions of customers around the world. All games offer a chat feature that allows players to communicate with each other in real time. Messages can be typed in more than 20 languages and are translated in real time using the Cloud Translation API. You have been asked to build an ML system to moderate the chat in real time while assuring that the performance is uniform across the various languages and without changing the serving infrastructure. You trained your first model using an in-house word2vec model for embedding the chat messages translated by the Cloud Translation API. However, the model has significant differences in performance across the different languages. How should you improve it?

   • AAdd a regularization term such as the Min-Diff algorithm to the loss function.
   • BTrain a classifier using the chat messages in their original language.
   • CReplace the in-house word2vec with GPT-3 or T5.
   • DRemove moderation for languages for which the false positive rate is too high.

   Show answer

   Answer: B

   Explanation

   Answer B Since the performance of the model varies significantly across different languages, it suggests that the translation process might have introduced some noise in the chat messages, making it difficult for the model to generalize across languages. One way to address this issue is to train a classifier using the chat messages in their original language.

   Source: Adapted from community discussion — TNT87

   Next question

8. Question 8 · 364 in bank

   You work for a gaming company that develops massively multiplayer online (MMO) games. You built a TensorFlow model that predicts whether players will make in-app purchases of more than $10 in the next two weeks. The model’s predictions will be used to adapt each user’s game experience. User data is stored in BigQuery. How should you serve your model while optimizing cost, user experience, and ease of management?

   • AImport the model into BigQuery ML. Make predictions using batch reading data from BigQuery, and push the data to Cloud SQL
   • BDeploy the model to Vertex AI Prediction. Make predictions using batch reading data from Cloud Bigtable, and push the data to Cloud SQL.
   • CEmbed the model in the mobile application. Make predictions after every in-app purchase event is published in Pub/Sub, and push the data to Cloud SQL.
   • DEmbed the model in the streaming Dataflow pipeline. Make predictions after every in-app purchase event is published in Pub/Sub, and push the data to Cloud SQL.

   Show answer

   Answer: A

   Explanation

   Option C, embedding the model in the mobile application, can increase the size of the application and may not be suitable for real-time prediction.

   Source: Adapted from community discussion — tavva_prudhvi

   References

   • Google Cloud documentation

   Next question

9. Question 9 · 364 in bank

   You are building a linear regression model on BigQuery ML to predict a customer’s likelihood of purchasing your company’s products. Your model uses a city name variable as a key predictive component. In order to train and serve the model, your data must be organized in columns. You want to prepare your data using the least amount of coding while maintaining the predictable variables. What should you do?

   • AUse TensorFlow to create a categorical variable with a vocabulary list. Create the vocabulary file, and upload it as part of your model to BigQuery ML.
   • BCreate a new view with BigQuery that does not include a column with city information
   • CUse Cloud Data Fusion to assign each city to a region labeled as 1, 2, 3, 4, or 5, and then use that number to represent the city in the model.
   • DUse Dataprep to transform the state column using a one-hot encoding method, and make each city a column with binary values.

   Show answer

   Answer: D

   Explanation

   A. Using TensorFlow: This is an overkill for this scenario. BigQuery ML can handle one-hot encoding natively within Dataprep. B. Excluding City Information: This removes a potentially important predictive variable, reducing model accuracy. C. Assigning Region Labels: This approach loses granularity and might not capture the specific variations between cities.

   Source: Adapted from community discussion — fitri001

   References

   • Google Cloud documentation

   Next question

10. Question 10 · 364 in bank

    You are an ML engineer at a bank that has a mobile application. Management has asked you to build an ML-based biometric authentication for the app that verifies a customer’s identity based on their fingerprint. Fingerprints are considered highly sensitive personal information and cannot be downloaded and stored into the bank databases. Which learning strategy should you recommend to train and deploy this ML mode?

    • AData Loss Prevention API
    • BFederated learning
    • CMD5 to encrypt data
    • DDifferential privacy

    Show answer

    Answer: B

    Explanation

    B With federated learning, all the data is collected, and the model is trained with algorithms across multiple decentralized edge devices such as cell phones or websites, without exchanging them. (Journey to Become a Google Cloud Machine Learning Engineer: Build the mind and hand of a Google Certified ML professional)

    Source: Adapted from community discussion — hiromi

    Next question

Keyboard:J Kmove between questionsA–Dselect an answerRshow or hide the answer← →change page

PMLE practice, question 1. 364 questions available.