CS3491 Artificial Intelligence and Machine Learning Previous Year Question Papers - Anna University
Access Anna University Artificial Intelligence and Machine Learning (CS3491) previous year question papers on LearnSkart for smarter semester exam preparation. This Anna University PYQ page offers year-wise Anna University exam papers aligned with Regulation 2021, so students can understand recurring questions, important units, and expected marking schemes. You can view every CS3491 Artificial Intelligence and Machine Learning question paper online and use free PDF download options for focused revision before internal and semester exams.
2024
-
2024 - CSE-AM-2024-CS 3491-Artificial Intelligence and Machine Learning-451033780-80904.pdf
-
2024 - CSE-ND-2024-CS 3491-Artificial intelligence and Machine Learning -131223726-20250604161745 (13).pdf
2023
-
2023 - CSE-AM-2023-CS 3491-Artificial intelligence and machine learning-594256739-AM23C (9).pdf
-
2023 - CSE-ND-2023-CS 3491-Artificial Intelligence and Machine Learning-769553624-20871.pdf
Important Questions - CS3491 Artificial Intelligence and Machine Learning
UNIT I: PROBLEM SOLVING
Part A (2 Marks)
- Define Artificial Intelligence and list its applications.
- What are the components of a well-defined problem?
- Differentiate Informed and Uninformed search.
- What is a Heuristic function?
Part B (13/15 Marks)
- Explain A* Search algorithm with an example and discuss optimality.
- Compare BFS and DFS (time and space complexity).
- Explain Constraint Satisfaction Problems (Map Coloring / 8-Queens).
- Discuss Hill Climbing and its drawbacks (Local maxima, plateaus).
UNIT II: KNOWLEDGE REPRESENTATION AND REASONING
Part A (2 Marks)
- What is First Order Logic (FOL)?
- Define Unification and Resolution.
- What is a Hidden Markov Model (HMM)?
- State Bayes' Rule and its significance.
Part B (13/15 Marks)
- Explain Resolution in Predicate Logic with proof.
- Discuss Bayesian Networks for uncertain knowledge.
- Compare Forward and Backward Chaining.
- Explain Exact Inference using Variable Elimination.
UNIT III: MACHINE LEARNING FUNDAMENTALS
Part A (2 Marks)
- Differentiate Supervised and Unsupervised learning.
- What is Hypothesis Space?
- Define Overfitting and Underfitting.
- What is Regularization?
Part B (13/15 Marks)
- Explain Version Space and Candidate Elimination Algorithm.
- Discuss Inductive Learning and bias.
- Explain Linear Regression with Gradient Descent.
- Describe PAC Learning framework.
UNIT IV: SUPERVISED LEARNING
Part A (2 Marks)
- What is a Decision Tree?
- Role of Entropy in selecting root node.
- Define Support Vectors in SVM.
- What is Kernel Trick?
Part B (13/15 Marks)
- Explain ID3 algorithm using Information Gain.
- Discuss K-Nearest Neighbors (KNN) with example.
- Explain SVM (linear & non-linear cases).
- Describe Neural Networks and Backpropagation.
UNIT V: UNSUPERVISED LEARNING AND ADVANCED TOPICS
Part A (2 Marks)
- What is Clustering?
- Define Principal Component Analysis (PCA).
- What is a Centroid in K-Means?
- Differentiate Hard and Soft Clustering.
Part B (13/15 Marks)
- Explain K-Means algorithm with numerical example.
- Discuss Hierarchical Clustering methods.
- Explain PCA for dimensionality reduction.
- Describe Reinforcement Learning and Q-Learning.
Most Repeated / High-Weight Questions
A* search and problem solving, knowledge representation and reasoning, decision trees with entropy calculation, support vector machines, neural networks and backpropagation, K-means clustering, PCA for dimensionality reduction.
Additional Resources
How to Use These Question Papers
- Problem Solving First: Master search algorithms (Unit I) with heuristic design and problem formulation. These are foundation concepts for understanding AI approaches throughout the course.
- Knowledge Representation Deep Dive: Practice FOL, unification, and resolution proofs (Unit II). Understand Bayesian networks and probabilistic reasoning for reasoning under uncertainty.
- ML Theory and Practice: Study supervised learning algorithms (Unit IV) with mathematical foundations. Practice decision tree construction using information gain, SVM kernel selection, and neural network backpropagation calculations.
- Clustering and Dimensionality Reduction: Implement K-means with manual centroid calculations and trace through iterations. Understand PCA transformation and its applications for feature reduction.
- Time Management: Allocate 60-90 minutes per algorithm explanation; practice Part B solutions combining concepts from multiple units under timed conditions.
Frequently Asked Questions about CS3491 Artificial Intelligence and Machine Learning
Which AI/ML topics are most frequently tested in CS3491 exams?
Search algorithms with heuristics (Unit I), knowledge representation and reasoning (Unit II), supervised learning algorithms (Unit IV), and clustering (Unit V) together account for 70% of exam marks. Unit III covers ML fundamentals essential for all algorithms. Practice combining concepts from multiple units.
How should I approach A* search algorithm in CS3491?
Understand A* combines g(n) (cost from start) and h(n) (heuristic estimate to goal): f(n) = g(n) + h(n). Practice with state space problems: draw graph, select nodes with lowest f value, maintain open/closed lists. Trace manually to understand optimality and completeness. These search problems appear with 13-15 marks in Unit I.
What is the best strategy for decision tree questions in CS3491?
Understand information gain and entropy calculation for ID3 algorithm. Calculate entropy of parent and children, compute information gain for each attribute. Select attribute with maximum gain as root. Recursively build tree. Practice with datasets manually to understand splitting criteria. These tree construction problems appear regularly.
How can I master SVM concepts in CS3491?
Understand margin concept (distance between hyperplane and closest points). Learn kernel trick: transform non-linear problem to linear in higher dimensions. Recognize linear (separate hyperplane) vs non-linear (RBF kernel) cases. These SVM concepts appear with 13-15 marks in Unit IV with emphasis on intuitive understanding.
What should I know about K-Means clustering in CS3491?
Initialize centroids, assign points to nearest centroid, recalculate centroids iteratively until convergence. Practice with numerical examples: plot 2D data, manually trace through iterations. Understand convergence criteria and limitations (local optima, requires k specification). Compare with hierarchical clustering methods.
How should I approach neural networks and backpropagation in CS3491?
Understand network architecture (input, hidden, output layers) and forward propagation. Master backpropagation: compute error at output, propagate backward calculating gradients for weight updates. Practice with simple 3-layer network manually. Recognize activation functions (sigmoid, ReLU) and their role. These neural network questions appear in Unit IV.