PMLE practice questions 251–260

1. Question 251 · 364 in bank

   You are training a deep learning model for semantic image segmentation with reduced training time. While using a Deep Learning VM Image, you receive the following error: The resource 'projects/deeplearning-platforn/zones/europe-west4-c/acceleratorTypes/nvidia-tesla-k80' was not found. What should you do?

   • AEnsure that you have GPU quota in the selected region.
   • BEnsure that the required GPU is available in the selected region.
   • CEnsure that you have preemptible GPU quota in the selected region.
   • DEnsure that the selected GPU has enough GPU memory for the workload.

   Show answer

   Answer: B

   Explanation

   ANS: B https://cloud.google.com/deep-learning-vm/docs/troubleshooting#resource_not_found https://cloud.google.com/compute/docs/gpus/gpu-regions-zones Resource not found Symptom: - The resource 'projects/deeplearning-platform/zones/europe-west4-c/acceleratorTypes/nvidia-tesla-k80' was not found Problem: You are trying to create an instance with one or more GPUs in a region where GPUs are not available (for example, an instance with a K80 GPU in europe-west4-c). Solution: To determine which region has the required GPU, see GPUs on Compute Engine.

   Source: Adapted from community discussion — celia20200410

   References

   • Google Cloud documentation
   • Google Cloud documentation

   Next question

2. Question 252 · 364 in bank

   Your team is working on an NLP research project to predict political affiliation of authors based on articles they have written. You have a large training dataset that is structured like this:

   

   You followed the standard 80%-10%-10% data distribution across the training, testing, and evaluation subsets. How should you distribute the training examples across the train-test-eval subsets while maintaining the 80-10-10 proportion?

   • ADistribute texts randomly across the train-test-eval subsets: Train set: [TextA1, TextB2, ...] Test set: [TextA2, TextC1, TextD2, ...] Eval set: [TextB1, TextC2, TextD1, ...]
   • BDistribute authors randomly across the train-test-eval subsets: (*) Train set: [TextA1, TextA2, TextD1, TextD2, ...] Test set: [TextB1, TextB2, ...] Eval set: [TexC1,TextC2 ...]
   • CDistribute sentences randomly across the train-test-eval subsets: Train set: [SentenceA11, SentenceA21, SentenceB11, SentenceB21, SentenceC11, SentenceD21 ...] Test set: [SentenceA12, SentenceA22, SentenceB12, SentenceC22, SentenceC12, SentenceD22 ...] Eval set: [SentenceA13, SentenceA23, SentenceB13, SentenceC23, SentenceC13, SentenceD31 ...]
   • DDistribute paragraphs of texts (i.e., chunks of consecutive sentences) across the train-test-eval subsets: Train set: [SentenceA11, SentenceA12, SentenceD11, SentenceD12 ...] Test set: [SentenceA13, SentenceB13, SentenceB21, SentenceD23, SentenceC12, SentenceD13 ...] Eval set: [SentenceA11, SentenceA22, SentenceB13, SentenceD22, SentenceC23, SentenceD11 ...]

   Show answer

   Answer: B

   Explanation

   Exactly! I also consider is B Check this out! If we just put inside the Training set , Validation set and Test set , randomly Text, Paragraph or sentences the model will have the ability to learn specific qualities about The Author's use of language beyond just his own articles. Therefore the model will mixed up different opinions. Rather if we divided things up a the author level, so that given authors were only on the training data, or only in the test data or only in the validation data. The model will find more difficult to get a high accuracy on the test validation (What is correct and have more sense!). Because it will need to really focus in author by author articles rather than get a single political affiliation based on a bunch of mixed articles from different authors. https://developers.google.com/machine-learning/crash-course/18th-century-literature

   Source: Adapted from community discussion — jk73

   References

   • Google developer documentation
   • Google Cloud documentation

   Next question

3. Question 253 · 364 in bank

   Your team has been tasked with creating an ML solution in Google Cloud to classify support requests for one of your platforms. You analyzed the requirements and decided to use TensorFlow to build the classifier so that you have full control of the model's code, serving, and deployment. You will use Kubeflow pipelines for the ML platform. To save time, you want to build on existing resources and use managed services instead of building a completely new model. How should you build the classifier?

   • AUse the Natural Language API to classify support requests.
   • BUse AutoML Natural Language to build the support requests classifier.
   • CUse an established text classification model on AI Platform to perform transfer learning.
   • DUse an established text classification model on AI Platform as-is to classify support requests.

   Show answer

   Answer: C

   Explanation

   the model cannot work as-is as the classes to predict will likely not be the same; we need to use transfer learning to retrain the last layer and adapt it to the classes we need, hence C

   Source: Adapted from community discussion — ms_lemon

   References

   • Google Cloud documentation

   Next question

4. Question 254 · 364 in bank

   You recently joined a machine learning team that will soon release a new project. As a lead on the project, you are asked to determine the production readiness of the ML components. The team has already tested features and data, model development, and infrastructure. Which additional readiness check should you recommend to the team?

   • AEnsure that training is reproducible.
   • BEnsure that all hyperparameters are tuned.
   • CEnsure that model performance is monitored.
   • DEnsure that feature expectations are captured in the schema.

   Show answer

   Answer: C

   Explanation

   Hey! all guys A+B+D=The team has already tested features and data, model development, and infrastructure. we are about to go live with production. Monitoring readiness is the last thing to account for. It will be very rediculous if you launch model as production regardless of how we will have about monitoring. you will lauch model as production for while and will make plan to model performance monitoring later ??? you are too reckless. Pls . Read it carefully https://developers.google.com/machine-learning/testing-debugging/pipeline/production https://developers.google.com/machine-learning/testing-debugging/pipeline/overview#what-is-an-ml-pipeline. You Most guys prefer A : https://developers.google.com/machine-learning/testing-debugging/pipeline/deploying I think that it is all about model development prior to deploying .

   Source: Adapted from community discussion — John_Pongthorn

   References

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

   Next question

5. Question 255 · 364 in bank

   You work for a credit card company and have been asked to create a custom fraud detection model based on historical data using AutoML Tables. You need to prioritize detection of fraudulent transactions while minimizing false positives. Which optimization objective should you use when training the model?

   • AAn optimization objective that minimizes Log loss
   • BAn optimization objective that maximizes the Precision at a Recall value of 0.50
   • CAn optimization objective that maximizes the area under the precision-recall curve (AUC PR) value
   • DAn optimization objective that maximizes the area under the receiver operating characteristic curve (AUC ROC) value

   Show answer

   Answer: C

   Explanation

   Detection of fraudulent transactions seems to be imbalanced data. https://cloud.google.com/automl-tables/docs/train#opt-obj AUC ROC : Distinguish between classes. Default value for binary classification. AUC PR Optimize results for predictions for the less common class. it is straightforward to answer, you just have to capture key word to get the right way. (Almost banlanced Or Imbalanced) https://machinelearningmastery.com/roc-curves-and-precision-recall-curves-for-classification-in-python/ When to Use ROC vs. Precision-Recall Curves? Generally, the use of ROC curves and precision-recall curves are as follows: ROC curves should be used when there are roughly equal numbers of observations for each class. Precision-Recall curves should be used when there is a moderate to large class imbalance.

   Source: Adapted from community discussion — John_Pongthorn

   References

   • Google developer documentation
   • Google Cloud documentation

   Next question

6. Question 256 · 364 in bank

   Your company manages a video sharing website where users can watch and upload videos. You need to create an ML model to predict which newly uploaded videos will be the most popular so that those videos can be prioritized on your company's website. Which result should you use to determine whether the model is successful?

   • AThe model predicts videos as popular if the user who uploads them has over 10,000 likes.
   • BThe model predicts 97.5% of the most popular clickbait videos measured by number of clicks.
   • CThe model predicts 95% of the most popular videos measured by watch time within 30 days of being uploaded.
   • DThe Pearson correlation coefficient between the log-transformed number of views after 7 days and 30 days after publication is equal to 0.

   Show answer

   Answer: C

   Explanation

   Ans: C (See https://developers.google.com/machine-learning/problem-framing/framing#quantify-it; though it's just an example.) (A) The absolute number of likes shouldn't be used because no information about subscribers or visits to the website is provided. The number may vary. (B) Clickbait videos are a subset of uploaded videos. Using them is an improper criterion. (D) The coefficient should reach 1. (Ref:https://arxiv.org/pdf/1510.06223.pdf)

   Source: Adapted from community discussion — Paul_Dirac

   References

   • Google developer documentation

   Next question

7. Question 257 · 364 in bank

   You are working on a Neural Network-based project. The dataset provided to you has columns with different ranges. While preparing the data for model training, you discover that gradient optimization is having difficulty moving weights to a good solution. What should you do?

   • AUse feature construction to combine the strongest features.
   • BUse the representation transformation (normalization) technique.
   • CImprove the data cleaning step by removing features with missing values.
   • DChange the partitioning step to reduce the dimension of the test set and have a larger training set.

   Show answer

   Answer: B

   Explanation

   I also think it is B: "The presence of feature value X in the formula will affect the step size of the gradient descent. The difference in ranges of features will cause different step sizes for each feature. To ensure that the gradient descent moves smoothly towards the minima and that the steps for gradient descent are updated at the same rate for all the features, we scale the data before feeding it to the model."

   Source: Adapted from community discussion — pddddd

   References

   • Google developer documentation

   Next question

8. Question 258 · 364 in bank

   You work for an online retail company that is creating a visual search engine. You have set up an end-to-end ML pipeline on Google Cloud to classify whether an image contains your company's product. Expecting the release of new products in the near future, you configured a retraining functionality in the pipeline so that new data can be fed into your ML models. You also want to use AI Platform's continuous evaluation service to ensure that the models have high accuracy on your test dataset. What should you do?

   • AKeep the original test dataset unchanged even if newer products are incorporated into retraining.
   • BExtend your test dataset with images of the newer products when they are introduced to retraining.
   • CReplace your test dataset with images of the newer products when they are introduced to retraining.
   • DUpdate your test dataset with images of the newer products when your evaluation metrics drop below a pre-decided threshold.

   Show answer

   Answer: B

   Explanation

   I think B is the right answer. A: Doesn't make sense. If you don't use the new product, it becomes useless. C: Conventional products are also necessary as data. D: I don't understand the need to wait until the threshold is exceeded.

   Source: Adapted from community discussion — esuaaaa

   Next question

9. Question 259 · 364 in bank

   You are using transfer learning to train an image classifier based on a pre-trained EfficientNet model. Your training dataset has 20,000 images. You plan to retrain the model once per day. You need to minimize the cost of infrastructure. What platform components and configuration environment should you use?

   • AA Deep Learning VM with 4 V100 GPUs and local storage.
   • BA Deep Learning VM with 4 V100 GPUs and Cloud Storage.
   • CA Google Kubernetes Engine cluster with a V100 GPU Node Pool and an NFS Server
   • DAn AI Platform Training job using a custom scale tier with 4 V100 GPUs and Cloud Storage

   Show answer

   Answer: D

   Explanation

   ans: D A, C => local storage, NFS... discarded. Google encourages you to use Cloud Storage. B => could do the job, but here I would focus on the "daily training" thing, because Vertex AI Training jobs are better for this. Also I think that Google usually encourages to use Vertex AI over VMs.

   Source: Adapted from community discussion — wish0035

   Next question

10. Question 260 · 364 in bank

    While conducting an exploratory analysis of a dataset, you discover that categorical feature A has substantial predictive power, but it is sometimes missing. What should you do?

    • ADrop feature A if more than 15% of values are missing. Otherwise, use feature A as-is.
    • BCompute the mode of feature A and then use it to replace the missing values in feature A.
    • CReplace the missing values with the values of the feature with the highest Pearson correlation with feature A.
    • DAdd an additional class to categorical feature A for missing values. Create a new binary feature that indicates whether feature A is missing.

    Show answer

    Answer: D

    Explanation

    ans: D A => no, you don't want to drop a feature with high prediction power. B => i think this could confuse the model... a better solution could be to fill missing values using an algorithm like Expectation Maximization, but using the mode i think is a bad idea in this case, because if you have a significant number of missing values (for example >10%) this would modify the "predictive power". you don't want to lose predictive power of a feature, just guide the model to learn when to use that feature and when to ignore it. C => this doesn't make any sense for me. not sure what i would do that. D => i think this could be a really good approach, and i'm pretty sure it would work pretty well a lot of models. the model would learn that when "is_available_feat_A" == True, then it would use the feature A, but whenever it is missing then it would try to use other features.

    Source: Adapted from community discussion — wish0035

    References

    • Google developer documentation

    Next question

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

PMLE practice, question 251. 364 questions available.