CS 7643 Deep Learning | Comprehensive
Practice Exam 250 Questions | Answers
with Rationales
Updated for Current Syllabus
Instructions
This practice exam is designed to mirror the scope and difficulty
of assessments in Georgia Tech's CS 7643 Deep Learning course.
Questions are organized by topic module and include a mix of
conceptual, mathematical, and application-oriented items. Each
question is followed by the correct Answer and a detailed
Rationale explaining the underlying principle.
Topics Covered:
• Module 1: Background & Machine Learning Fundamentals (Q1–
30)
• Module 2: Optimization & Computation Graphs (Q31–60)
• Module 3: Neural Networks & Activation Functions (Q61–90)
• Module 4: Convolutional Neural Networks (Q91–130)
• Module 5: Training, Regularization & Transfer Learning (Q131–
170)
• Module 6: Sequence Modeling (RNNs, LSTMs, GRUs) (Q171–200)
• Module 7: Attention & Transformers (Q201–225)
• Module 8: Generative Models (GANs, VAEs, Diffusion) (Q226–250)
,Module 1: Background & Machine Learning Fundamentals
Q1. What distinguishes deep learning from traditional
machine learning?
A. Deep learning requires labeled data exclusively
B. Deep learning automatically learns hierarchical feature
representations from raw data
C. Deep learning uses only linear models
D. Deep learning cannot be applied to image data
Answer: B
Rationale: Deep learning focuses on learning complex,
hierarchical feature representations directly from raw data
through multiple layers of abstraction, whereas traditional
machine learning often relies on manually engineered features.
Deep learning can use labeled, unlabeled, or self-supervised data
and is highly effective for images, text, and audio.
Q2. Which of the following is NOT a typical component of a
deep learning model?
A. Linear layers
B. Convolutional layers
C. Hard-coded decision trees
D. Pooling layers
Answer: C
Rationale: Deep learning models are composed of differentiable
modules such as linear, convolutional, and pooling layers that are
learned from data. Hard-coded decision trees are not
differentiable and are not typical components of deep neural
networks.
,Q3. What is the primary objective of representation learning
in deep learning?
A. To manually engineer features for each task
B. To automatically discover useful features/representations for a
task from raw data
C. To reduce the number of parameters in a model
D. To eliminate the need for optimization algorithms
Answer: B
Rationale: Representation learning is the core objective of deep
learning—automatically discovering useful features or
representations for a task directly from raw data, without manual
feature engineering.
Q4. Which mathematical prerequisites are required for CS
7643 at Georgia Tech?
A. Introduction to Psychology
B. Linear algebra, calculus (partial derivatives),
probability/statistics, and an introductory machine learning
course
C. Web development experience
D. No prerequisites are required
Answer: B
Rationale: The course requires a strong mathematical
background including linear algebra, multivariate calculus, and
probability/statistics, plus at least an introductory machine
learning course (e.g., CS 7641). Self-study does not satisfy this
requirement.
, Q5. Which programming language is primarily used for
hands-on assignments in CS 7643?
A. Java
B. C++
C. Python
D. R
Answer: C
Rationale: Python is used throughout the course, first with
NumPy for foundational implementations and later with PyTorch
for deep learning.
Q6. What is the role of PyTorch in CS 7643?
A. It is used as a deep learning library for implementing and
training neural networks
B. It is used only for data visualization
C. It replaces the need for understanding backpropagation
D. It is a database management system
Answer: A
Rationale: PyTorch is a popular deep learning library used in the
course to implement neural network modules, perform automatic
differentiation, and train models efficiently.
Q7. What is a parametric model in the context of machine
learning?
Practice Exam 250 Questions | Answers
with Rationales
Updated for Current Syllabus
Instructions
This practice exam is designed to mirror the scope and difficulty
of assessments in Georgia Tech's CS 7643 Deep Learning course.
Questions are organized by topic module and include a mix of
conceptual, mathematical, and application-oriented items. Each
question is followed by the correct Answer and a detailed
Rationale explaining the underlying principle.
Topics Covered:
• Module 1: Background & Machine Learning Fundamentals (Q1–
30)
• Module 2: Optimization & Computation Graphs (Q31–60)
• Module 3: Neural Networks & Activation Functions (Q61–90)
• Module 4: Convolutional Neural Networks (Q91–130)
• Module 5: Training, Regularization & Transfer Learning (Q131–
170)
• Module 6: Sequence Modeling (RNNs, LSTMs, GRUs) (Q171–200)
• Module 7: Attention & Transformers (Q201–225)
• Module 8: Generative Models (GANs, VAEs, Diffusion) (Q226–250)
,Module 1: Background & Machine Learning Fundamentals
Q1. What distinguishes deep learning from traditional
machine learning?
A. Deep learning requires labeled data exclusively
B. Deep learning automatically learns hierarchical feature
representations from raw data
C. Deep learning uses only linear models
D. Deep learning cannot be applied to image data
Answer: B
Rationale: Deep learning focuses on learning complex,
hierarchical feature representations directly from raw data
through multiple layers of abstraction, whereas traditional
machine learning often relies on manually engineered features.
Deep learning can use labeled, unlabeled, or self-supervised data
and is highly effective for images, text, and audio.
Q2. Which of the following is NOT a typical component of a
deep learning model?
A. Linear layers
B. Convolutional layers
C. Hard-coded decision trees
D. Pooling layers
Answer: C
Rationale: Deep learning models are composed of differentiable
modules such as linear, convolutional, and pooling layers that are
learned from data. Hard-coded decision trees are not
differentiable and are not typical components of deep neural
networks.
,Q3. What is the primary objective of representation learning
in deep learning?
A. To manually engineer features for each task
B. To automatically discover useful features/representations for a
task from raw data
C. To reduce the number of parameters in a model
D. To eliminate the need for optimization algorithms
Answer: B
Rationale: Representation learning is the core objective of deep
learning—automatically discovering useful features or
representations for a task directly from raw data, without manual
feature engineering.
Q4. Which mathematical prerequisites are required for CS
7643 at Georgia Tech?
A. Introduction to Psychology
B. Linear algebra, calculus (partial derivatives),
probability/statistics, and an introductory machine learning
course
C. Web development experience
D. No prerequisites are required
Answer: B
Rationale: The course requires a strong mathematical
background including linear algebra, multivariate calculus, and
probability/statistics, plus at least an introductory machine
learning course (e.g., CS 7641). Self-study does not satisfy this
requirement.
, Q5. Which programming language is primarily used for
hands-on assignments in CS 7643?
A. Java
B. C++
C. Python
D. R
Answer: C
Rationale: Python is used throughout the course, first with
NumPy for foundational implementations and later with PyTorch
for deep learning.
Q6. What is the role of PyTorch in CS 7643?
A. It is used as a deep learning library for implementing and
training neural networks
B. It is used only for data visualization
C. It replaces the need for understanding backpropagation
D. It is a database management system
Answer: A
Rationale: PyTorch is a popular deep learning library used in the
course to implement neural network modules, perform automatic
differentiation, and train models efficiently.
Q7. What is a parametric model in the context of machine
learning?
