45 MCQs on Deep Learning and Related Topics.docx

3000

1. 45 MCQs on Deep Learning and Related Topics
Section 1: Basics of Deep Learning (ANN, RNN, CNN)
1. Which of the following is a building block of Artificial Neural Networks?
a) Nodes
b) Layers
c) Weights
d) All of the above
2. What is the purpose of an activation function in a neural network?
a) To control the weight initialization
b) To introduce non-linearity
c) To reduce overfitting
d) To ensure gradient flow
3. Which algorithm is commonly used to optimize the weights in a neural network?
a) Gradient Descent
b) Backpropagation
c) Genetic Algorithm
d) k-Means Clustering
4. What is a Recurrent Neural Network (RNN) primarily used for?
a) Image processing
b) Temporal sequence data
c) Data clustering
d) Data summarization
5. Which type of RNN solves the vanishing gradient problem?
a) Simple RNN
b) LSTM
c) GRU
d) Both b and c
6. In CNNs, which operation is used to reduce the spatial dimensions of feature maps?
a) Convolution

2. b) Pooling
c) Fully connected layers
d) ReLU activation
7. What does the kernel or filter in a CNN do?
a) Aggregates features
b) Detects specific patterns
c) Normalizes the data
d) Introduces non-linearity
8. What is the main purpose of dropout in a neural network?
a) To increase the number of neurons
b) To prevent overfitting
c) To make the model faster
d) To enhance interpretability
9. Which component of CNN detects edges in images?
a) Activation function
b) Filters
c) Fully connected layers
d) Optimizer
10. The output of a softmax function represents:
a) Probabilities of classes
b) Raw predictions
c) Feature maps
d) Activation values
Section 2: Natural Language Processing (NLP)
11. Which of the following techniques is commonly used for word representation in NLP?
a) One-hot encoding
b) Word embeddings
c) Frequency distribution
d) Bag of Words

3. 12. What is the primary purpose of tokenization in NLP?
a) To compress data
b) To divide text into smaller units
c) To apply stemming
d) To vectorize data
13. Which model architecture is most commonly used for language translation tasks?
a) RNN
b) Transformer
c) CNN
d) Logistic Regression
14. What is the full form of BERT in NLP?
a) Bidirectional Encoder Representations from Transformers
b) Bi-directional Entity Recognition Transformer
c) Binary Encoding for Robust Text
d) Bidirectional Extensible Recurrent Transformer
15. Which of the following is an application of NLP?
a) Chatbots
b) Text summarization
c) Sentiment analysis
d) All of the above
16. What is the purpose of stemming in NLP?
a) To remove stop words
b) To normalize words to their root form
c) To vectorize the text
d) To split sentences
17. Named Entity Recognition (NER) identifies:
a) Parts of speech
b) Specific entities like names, locations, or dates
c) Synonyms of words
d) Word frequencies
18. Which dataset is popular for NLP tasks?

4. a) MNIST
b) COCO
c) IMDb reviews
d) CIFAR-10
19. Word2Vec generates:
a) One-hot encoded vectors
b) Dense vector embeddings
c) Bag of words representation
d) Sparse vector representations
20. What is BLEU score used for in NLP?
a) Evaluate translation quality
b) Detect sentence similarity
c) Measure text compression
d) Rank search engine results
Section 3: PySpark and Hadoop
21. What is PySpark used for?
a) Data visualization
b) Distributed data processing
c) Model evaluation
d) Neural network training
22. Which programming language is native to Hadoop?
a) Python
b) Java
c) Scala
d) R
23. HDFS stands for:
a) Hadoop Distributed File Storage
b) High Data File System
c) Hadoop Data Framework Storage
d) Hadoop Distributed File System

5. 24. In PySpark, RDD stands for:
a) Resilient Distributed Data
b) Randomized Distributed Dataset
c) Resilient Distributed Dataset
d) Random Data Distribution
25. What is the main purpose of Hadoop's MapReduce?
a) Data storage
b) Parallel processing of large datasets
c) Real-time data analysis
d) Visualization
26. Which is a core component of Hadoop?
a) Hive
b) HDFS
c) Apache Spark
d) TensorFlow
27. In PySpark, which function is used to apply transformations?
a) filter()
b) map()
c) reduce()
d) All of the above
28. What is a Spark DataFrame?
a) A distributed collection of data organized into named columns
b) A single-node database
c) A data visualization framework
d) A file format
29. Which scheduling mode does Hadoop use?
a) Static scheduling
b) Dynamic scheduling
c) Batch scheduling
d) FIFO scheduling

6. 30. PySpark supports the processing of:
a) Structured data
b) Semi-structured data
c) Unstructured data
d) All of the above
Section 4: Object Detection and Big Data
31. Which of the following is NOT a popular object detection model?
a) YOLO
b) SSD
c) ResNet
d) Faster R-CNN
32. What is the main task of IoU (Intersection over Union) in object detection?
a) Classify objects
b) Evaluate localization accuracy
c) Compute feature maps
d) Optimize weights
33. What does NMS (Non-Maximum Suppression) do in object detection?
a) Combines overlapping boxes
b) Removes redundant bounding boxes
c) Reduces the model size
d) Improves model accuracy
34. What is Big Data characterized by?
a) Volume
b) Velocity
c) Variety
d) All of the above
35. Which of the following tools is best suited for real-time Big Data analysis?
a) Hadoop
b) Spark Streaming
c) Hive
d) Sqoop

7. 36. Apache Hadoop is primarily designed for:
a) Real-time data analytics
b) Batch processing of large datasets
c) Interactive data querying
d) Small-scale data storage
37. Which algorithm is used for clustering Big Data?
a) K-Means
b) Linear Regression
c) Decision Trees
d) SVM
38. What is a primary advantage of Big Data frameworks like Hadoop?
a) High scalability
b) Real-time performance
c) Low hardware cost
d) Single-threaded processing
39. YOLO in object detection stands for:
a) You Only Look Once
b) Your Object Localization Output
c) Yearly Object Label Optimization
d) None of the above
40. Which type of neural network is most suitable for object detection tasks?
a) RNN
b) CNN
c) Transformer
d) MLP
Section 5: Mixed Concepts
41. What is the main advantage of using deep learning for Big Data?
a) Faster computation
b) Better feature extraction
c) Simpler architecture
d) Reduced memory requirements

8. 42. Which is an example of unstructured Big Data?
a) Emails
b) Images
c) Videos
d) All of the above
43. Hadoop's NameNode is responsible for:
a) Data storage
b) Metadata management
c) Data processing
d) Network optimization
44. What is the role of batch normalization in a neural network?
a) Reduce training time
b) Prevent overfitting
c) Normalize layer inputs
d) All of the above
45. Which deep learning model is best suited for time-series forecasting?
a) RNN
b) CNN
c) Autoencoder
d) GAN