最新的高品質 NCP-ADS 考試練習題庫

You are designing an accelerated ETL pipeline to process large-scale datasets in a data science workflow.
Which of the following are key considerations when selecting the right tools and methods for implementing this pipeline? (Select two)

• A. Leveraging parallel processing and distributed computing frameworks like Apache Spark to speed up the transformation phase.
• B. Using GPU-accelerated libraries such as RAPIDS for data transformation to enhance processing speed.
• C. Ensuring the ETL pipeline uses only batch processing for data ingestion.
• D. Relying on traditional single-threaded processing for the extraction phase to reduce complexity.
• E. Using a single storage location for both raw and transformed data to simplify the workflow.

You are training a deep learning model for image classification and want to optimize its hyperparameters, including learning rate, batch size, and number of layers.
Which of the following techniques is the most effective for efficiently searching through a high- dimensional hyperparameter space?

• A. Bayesian Optimization
• B. Random Search
• C. Gradient Descent
• D. Grid Search

A team is processing large-scale tabular data using cuDF and cuML on NVIDIA GPUs but is facing performance degradation.
Which of the following techniques would be the most effective in identifying and resolving bottlenecks in the pipeline?

• A. Increase GPU clock speed manually to force higher processing power.
• B. Convert all datasets into pandas DataFrames for initial processing before moving to cuDF.
• C. Reduce the dataset size to a smaller sample to speed up processing.
• D. Use nvprof or nsight compute to analyze kernel execution time and memory transfer efficiency.

A data scientist wants to compare the performance of two different GPU-accelerated data science frameworks, NVIDIA RAPIDS (cuDF, cuML) and TensorFlow, for a tabular data classification task.
Which of the following approaches would be the best practice for designing an unbiased and effective benchmark?

• A. Run all benchmarks on a CPU to ensure fairness across frameworks.
• B. Measure execution time and memory usage for each framework using NVIDIA Nsight Systems (nsys).
• C. Ignore preprocessing and focus only on model training speed when comparing performance.
• D. Use TensorFlow's built-in training time metrics without comparing equivalent RAPIDS-based operations.

You are working on a financial dataset that tracks stock prices over time, and you need to detect anomalies such as sudden spikes or drops using NVIDIA technologies.
Which of the following approaches would be the most effective for anomaly detection in a time-series dataset using NVIDIA's RAPIDS AI and TensorRT?

• A. Apply a traditional rule-based thresholding method using pandas and NumPy for detecting sudden spikes in stock prices.
• B. Perform anomaly detection by applying DBSCAN clustering with RAPIDS cuML without any feature engineering.
• C. Use RAPIDS cuML's Isolation Forest for anomaly detection and deploy it with NVIDIA Triton Inference Server.
• D. Use traditional ARIMA modeling with RAPIDS cuML to classify anomalies based on residual analysis.