MLS-C01 Mock Exam 71 — 25 Questions

Question 1

A company uses Amazon Kinesis Data Analytics for Apache Flink to process streaming data. The application reads from a Kinesis data stream and writes results to a sink. The application is failing with an 'OutOfMemoryError'. The application has parallelism set to 4 and uses 1 Kinesis Processing Unit (KPU). What is the MOST likely cause and solution?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

The heap memory per operator is too low; increase parallelism to 8.. With parallelism set to 4 but only 1 KPU, each operator slot receives a fraction of the available heap memory, leading to an OutOfMemoryError. Increasing parallelism to 8 distributes the workload across more slots, but more importantly, it forces Kinesis Data Analytics to allocate additional KPUs (each KPU provides 4 GB of memory), thereby increasing the total heap memory available to the application.

Answer

The application is using too many operators; reduce parallelism to 2.

Answer

The checkpoint interval is too short; increase it to 5 minutes.

Answer

The buffer timeout is too high; reduce it to 50 ms.

Question 2

A research team is developing a deep learning model to classify medical images into 10 disease categories. They have a dataset of 50,000 labeled images, but the class distribution is highly imbalanced: the most common class has 20,000 images, while the rarest class has only 200 images. To address this, they apply data augmentation (random rotations, flips, and brightness adjustments) to the minority classes until each class has 20,000 images. They then train a convolutional neural network (CNN) from scratch using cross-entropy loss. The model achieves 95% overall accuracy but only 30% recall on the rarest class. Which change is MOST likely to improve recall on the rarest class without significantly reducing overall accuracy?

Accepted Answer

Replace cross-entropy loss with focal loss. Focal loss is specifically designed to address class imbalance by down-weighting the loss contribution from well-classified examples (majority classes) and focusing training on hard, misclassified examples (minority classes). This directly improves recall on the rarest class, while cross-entropy loss treats all classes equally, causing the model to be biased toward the majority classes.

Answer

Increase dropout rate from 0.2 to 0.5 to reduce overfitting

Answer

Switch from Adam optimizer to SGD with momentum

Answer

Reduce the batch size from 64 to 16 to increase stochasticity

Question 3

A machine learning engineer needs to deploy a model that requires custom inference code with dependencies. Which SageMaker deployment option should be used?

Accepted Answer

Create a custom Docker container and deploy to SageMaker endpoint.. When a model requires custom inference code with dependencies, the only way to fully control the runtime environment, libraries, and inference logic is to package everything into a custom Docker container. SageMaker endpoints can then deploy this container, allowing the engineer to specify the exact inference script and dependencies (e.g., via a Dockerfile and a requirements.txt). This approach supports any framework or custom logic that built-in algorithms cannot provide.

Answer

Use a SageMaker notebook instance as an endpoint.

Answer

Use a built-in SageMaker algorithm.

Answer

Use a SageMaker batch transform job.

Question 4

Which THREE techniques are commonly used in exploratory data analysis to understand the relationships between features and the target variable? (Select THREE.)

Accepted Answer

Use box plots to compare feature distributions across target classes.. Options A, C, and D are correct. Box plots (A) are useful for comparing feature distributions across different target classes, revealing differences that may indicate predictive power. Scatter plots or pair plots (D) allow visual inspection of relationships between features and the target, highlighting patterns, clusters, or outliers. A correlation matrix (C) quantifies linear relationships between features and the target variable, helping identify strongly correlated features. B is incorrect because K-means clustering is an unsupervised technique used for grouping data, not for understanding feature-target relationships. E is incorrect because PCA is a dimensionality reduction technique, not a direct method for analyzing relationships between features and a target variable.

Answer

Perform K-means clustering on the features.

Answer

Apply Principal Component Analysis (PCA) to reduce dimensions.

Question 5

A data scientist is training a linear regression model on a dataset with 50 features. After training, they notice that the model performs well on training data but poorly on test data. They suspect overfitting. Which action should they take to reduce overfitting?

Accepted Answer

Add L2 regularization (Ridge regression). L2 regularization (Ridge regression) adds a penalty term proportional to the square of the magnitude of the coefficients to the loss function. This discourages the model from fitting the noise in the training data by shrinking the weights, which reduces variance and mitigates overfitting, improving generalization to the test data.

Answer

Use a larger learning rate

Answer

Add more features to the model

Answer

Increase the number of training epochs

Question 6

Which THREE are valid considerations when deploying a large deep learning model (10 GB) on a SageMaker endpoint? (Choose 3.)

Accepted Answer

Enable SageMaker Data Compression for network transfer.. SageMaker Data Compression uses HTTP compression (e.g., gzip) to reduce the payload size during network transfer between the client and endpoint, which is critical for a 10 GB model to minimize latency and bandwidth consumption. This is especially beneficial when the model is large and inference requests involve substantial input or output data.

Answer

Use SageMaker Serverless Inference to avoid managing instances.

Answer

Attach Elastic Inference accelerators.

Question 7

A data scientist is exploring a dataset and wants to understand the distribution of a continuous feature. Which visualization is most appropriate for identifying skewness and potential outliers?

Accepted Answer

Box plot. (bar chart) is wrong because bar charts are for categorical data, not for showing distribution of a continuous feature. Option B (scatter plot) is wrong because scatter plots show relationships between two variables, not distribution. Option C is correct because a box plot explicitly shows median, quartiles, and outliers. Option D is wrong because heatmaps show correlations, not distribution.

Answer

Bar chart

Answer

Scatter plot

Answer

Heatmap

Question 8

Refer to the exhibit. A data scientist checks the status of a SageMaker endpoint and sees the output above. What does this indicate?

Accepted Answer

The endpoint is scaling up to meet desired capacity. The endpoint shows InService status with current instance count (2) less than desired count (5), indicating it is scaling up to meet desired capacity. Option D correctly describes this state. Option A is wrong because the status is InService, not OutOfService. Option B is incorrect because the endpoint is not at full capacity; it is under-provisioned. Option C is wrong because the endpoint is operational, not failed.

Answer

The endpoint has failed

Answer

The endpoint is running at full capacity

Answer

The endpoint is out of service

Question 9

Which THREE techniques are commonly used for feature engineering in exploratory data analysis? (Select THREE.)

Accepted Answer

Extracting date/time components like day of week or hour.. Extracting date/time components such as day of week, hour, or month from a timestamp is a standard feature engineering technique. It transforms a single datetime column into multiple categorical or cyclical features that can reveal temporal patterns like weekly seasonality or peak hours, which are often critical for time-series models.

Answer

Using principal component analysis (PCA) to create new features.

Answer

Applying one-hot encoding to numerical features.

Question 10

A company uses Amazon SageMaker to deploy a model for real-time inference. The model is a linear regression model that was trained using the SageMaker built-in Linear Learner algorithm. The endpoint is configured with an ml.m5.large instance. After deployment, the company notices that the endpoint returns incorrect predictions. The training data was normalized, but the inference requests send raw feature values without normalization. What should the company do to fix the issue?

Accepted Answer

Create a SageMaker inference pipeline that includes a preprocessing step to normalize the input data before passing it to the model.. The model was trained on normalized data, so it expects normalized input at inference time. Raw feature values will produce incorrect predictions because the model's coefficients are based on normalized data. The correct solution is to create a SageMaker inference pipeline that includes a preprocessing step (e.g., using a scikit-learn container) to normalize the input data before passing it to the model container. Option A (retrain with raw data) would require retraining and might degrade performance if normalization was necessary for convergence. Option B (changing instance type) does not address the data mismatch. Option D (batch transform job) is for batch inference, not real-time inference via an endpoint.

Answer

Retrain the model using raw data without normalization.

Answer

Change the endpoint instance type to a GPU instance to handle the raw data.

Answer

Use a batch transform job to preprocess the data before sending it to the endpoint.

Question 11

A data scientist is training a binary classifier on an imbalanced dataset where the positive class represents 1% of the data. The model is evaluated using accuracy, but the accuracy is 99% even though the model predicts all instances as negative. Which metric should the data scientist use to properly evaluate the model?

Accepted Answer

F1 score. The F1 score is the harmonic mean of precision and recall, making it robust to class imbalance. With 99% negative instances, accuracy is misleadingly high even if the model never predicts the positive class. F1 captures both false positives and false negatives, providing a balanced evaluation of the minority class performance.

Answer

Root mean squared error (RMSE)

Answer

Mean squared error (MSE)

Answer

Accuracy

Question 12

Match each hyperparameter tuning strategy to its description.

Accepted Answer

Grid Search: Exhaustively tries all combinations of specified hyperparameter values.. The correct matches are: Grid Search exhaustively searches all combinations; Random Search samples randomly; Bayesian Optimization uses a probabilistic model. Common confusions involve swapping the descriptions of Grid Search and Random Search, or confusing Random Search with Bayesian Optimization.

Answer

Grid Search: Randomly samples hyperparameter combinations from a specified distribution.

Answer

Random Search: Uses a probabilistic model to guide the search towards promising hyperparameters.

Question 13

A data science team is using Amazon SageMaker to train a model. The training job is failing with an 'OutOfMemory' error. The team is using a p3.2xlarge instance with 61 GB of memory. They need to resolve this issue as quickly as possible. Which action should they take?

Accepted Answer

Use a larger instance type, such as p3.8xlarge. The 'OutOfMemory' error indicates the training job requires more memory than the 61 GB available on the p3.2xlarge instance. The fastest resolution is to scale vertically by using a larger instance type, such as the p3.8xlarge, which provides 244 GB of memory. This directly addresses the memory shortage without requiring code changes or architectural modifications, minimizing downtime.

Answer

Reduce the batch size in the training script

Answer

Use a spot instance to save costs

Answer

Enable distributed training across multiple instances

Question 14

A company is deploying a real-time inference endpoint with SageMaker. The model is a large neural network that requires GPU acceleration. Which TWO configurations must be set?

Accepted Answer

Instance type with GPU. Deploying a real-time inference endpoint with a large neural network that requires GPU acceleration necessitates selecting an instance type with a GPU, such as the ml.p3 or ml.g4dn series, to provide the parallel processing power needed for low-latency inference. Without a GPU instance, the model would fall back to CPU, leading to unacceptable inference times for large neural networks.

Answer

Batch transform job

Answer

Production variant

Answer

Training container image

Question 15

A data scientist is using principal component analysis (PCA) for dimensionality reduction before training a classifier. The classifier's performance on the test set is poor. What is the most likely cause?

Accepted Answer

Too few principal components were retained, losing important information. C is correct because PCA is an unsupervised dimensionality reduction technique that projects data onto principal components capturing the maximum variance. If too few components are retained, the reduced representation may discard features that are critical for the classifier to distinguish between classes, leading to poor test performance due to underfitting.

Answer

The classifier is overfitting

Answer

The data was not scaled before applying PCA

Answer

Too many principal components were retained, including noise

Question 16

A company is deploying a real-time fraud detection system using a gradient boosting model on AWS SageMaker. The model uses 200 features and is trained on 50 GB of data. The inference latency requirement is under 10 ms per request. During load testing, the endpoint shows average latency of 15 ms. Which change is MOST likely to reduce latency below 10 ms?

Accepted Answer

Reduce the number of features to the top 50 based on feature importance. Reducing the number of features from 200 to the top 50 directly decreases the amount of data each inference request must process, which lowers both feature engineering overhead and model evaluation time. For gradient boosting models on SageMaker, fewer features mean fewer decision tree splits to traverse per prediction, which can significantly reduce latency without requiring hardware changes. This is the most direct and cost-effective way to meet the 10 ms requirement.

Answer

Switch to a GPU-based instance type

Answer

Increase the number of trees in the model

Answer

Use a larger batch size for inference

Question 17

A data scientist is analyzing a dataset of customer reviews. The dataset contains a text column 'review' and a numerical rating from 1 to 5. The data scientist wants to create features for sentiment analysis. Which THREE preprocessing steps should be applied to the text data before feature extraction? (Choose THREE.)

Accepted Answer

Apply stemming to reduce words to their root form.. Stemming reduces words to their root form (e.g., 'running' to 'run'), which consolidates variations of the same word and reduces feature dimensionality. This is a standard preprocessing step before feature extraction in NLP tasks like sentiment analysis, as it helps the model generalize across different word forms.

Answer

Standardize the text data using z-score normalization.

Answer

Tokenize the text into individual words.

Question 18

A team has deployed a SageMaker endpoint for a sentiment analysis model. The model was trained on text data from social media. After deployment, the team notices that the model's accuracy has dropped significantly after 3 months. Which action should the team take to detect and address this issue?

Accepted Answer

Enable SageMaker Model Monitor to detect data drift and trigger a retraining pipeline.. B is correct because SageMaker Model Monitor is specifically designed to detect data drift (changes in the input data distribution over time) and model drift (degradation in prediction quality). When a sentiment analysis model trained on social media text sees a drop in accuracy after months, it is likely due to shifts in language, slang, or topics. Model Monitor can continuously track the distribution of input features and predictions against a baseline, and when drift is detected, it can automatically trigger a retraining pipeline to update the model, directly addressing the root cause of the accuracy drop.

Answer

Use SageMaker A/B testing to compare with a new model.

Answer

Re-deploy the model using the same training script.

Answer

Create a CloudWatch alarm on invocation errors.

Question 19

A data engineer is setting up a data lake on Amazon S3 for a large retail company. The data includes customer transactions, inventory, and web logs. The company wants to use AWS Glue for ETL and Amazon Athena for ad-hoc queries. The data is partitioned by year, month, day, and hour. The engineer notices that Athena queries are slow and often scan large amounts of data even when only a specific hour is needed. The engineer has already enabled partitioning and used columnar formats like Parquet. What additional step should the engineer take to optimize query performance and reduce data scanned?

Accepted Answer

Use a coarser partition layout, such as partitioning only by date, and leverage Hive-style partitioning with AWS Glue Crawlers to avoid excessive small files.. Partitioning at a granularity of hour can result in a large number of small files, increasing metadata overhead and slowing query planning. By using a coarser partition layout (e.g., by date) and leveraging Hive-style partitioning with AWS Glue Crawlers, the number of partitions is reduced, which improves query performance and reduces the amount of data scanned. Option B is incorrect because converting Parquet to CSV would increase storage and scan costs due to lack of columnar compression and predicate pushdown. Option C is incorrect because S3 Select operates on a single object, not across partitions; it is not designed for optimizing Athena queries over many files. Option D is incorrect because increasing partition granularity (e.g., minute-level) would create even more small files, worsening the issue.

Answer

Convert the Parquet files to CSV format to reduce the overhead of columnar storage and improve compression.

Answer

Use S3 Select to push down filters to S3, reducing the amount of data scanned by Athena.

Answer

Increase the granularity of partitioning to include minute-level partitions to further limit data scanned.

Question 20

Which THREE techniques are effective for reducing overfitting in a deep neural network?

Accepted Answer

Early stopping. Dropout (C) randomly drops neurons during training, forcing the network to learn redundant representations and reducing overfitting. L2 regularization (E) adds a penalty on large weights, discouraging complex models. Early stopping (B) monitors validation loss and halts training before the model starts overfitting. Increasing model complexity (A) would worsen overfitting, and reducing training data (D) also increases overfitting risk due to less generalization. Thus, the correct techniques are B, C, and E.

Answer

Increasing model complexity

Answer

Reducing the amount of training data