Data Science and Artificial Intelligence

This course introduces students to data science and artificial intelligence. Students work with different types of data, form hypotheses and regression models, learn the foundations of intelligent systems, and study techniques for building them. The course also surveys AI areas such as natural language processing and robotics.

The course has two connected parts. The data science component introduces data formats, analysis tools in Python, visualization, data collection, hypothesis development, statistical inference, and regression.

The artificial intelligence component covers intelligent agents, search, constraint satisfaction, optimization, knowledge representation, logic, probabilistic reasoning, machine learning, neural networks, deep learning, natural language processing, speech recognition, and chatbots. It also introduces generative AI, computer vision, robotics, AI safety, ethics, and societal impact. Labs and assignments provide opportunities to apply the methods.

Work with different types of data.
Form hypotheses and regression models.
Understand core concepts of artificial intelligence and intelligent systems.
Learn techniques for building intelligent systems and associated data-processing methods.
Gain an overview of AI subfields, including natural language processing and robotics.

After completing this course, students should be able to:
1. Use the basic vocabulary of data science and explain core AI concepts and foundational techniques (Knowledge).
2. Visualize data, make inferences, and apply AI models to real-world problems (Apply).
3. Critique methods for drawing conclusions from data and evaluate intelligent systems (Evaluate).
4. Plan data visualization and analysis steps to gain insights from real-world data, and design and build a functional AI chatbot (Create).

Artificial Intelligence: A Modern Approach, Stuart Russell and Peter Norvig, 2015.
A First Course on Artificial Intelligence, Deepak Khemani, McGraw Hill India, 2013.
Introduction to Artificial Intelligence and Expert Systems, Dan W. Patterson, 1st edition, Prentice Hall of India, 2015.
Statistics: Unlocking the Power of Data, Robin H. Lock, Patti F. Lock, Kari L. Morgan, Eric F. Lock, and Dennis F. Lock, Wiley.
An Introduction to Statistical Learning with Applications in Python, James, Witten, Hastie, Tibshirani, and Taylor, Springer.

e-Resources
NPTEL: https://nptel.ac.in/courses/106102220
NPTEL: https://nptel.ac.in/courses/106105077
NPTEL: https://nptel.ac.in/courses/106105158
NPTEL: https://nptel.ac.in/courses/106106140
Harvard CS50 AI: https://cs50.harvard.edu/ai/2024/weeks/

Grading is relative for Freshmore courses.
Exams: 55% (final comprehensive exam 25%; two tests, 15% each).
Lab submissions: 20%.
Attendance: 10%.
Assignments: 15% (two assignments).

Week 1: Data handling, plotting, and visualization in Python.
Week 2: Data imputation and hypothesis testing review.
Weeks 3–4: Linear regression.
Week 5: Multiple and logistic regression.
Week 6: Decision trees.
Week 7: Intelligent agents, agent design, ethics, AI safety, and society.
Week 8: Uninformed search and problem solving.
Week 9: Informed search, heuristics, and optimization.
Week 10: Adversarial search, constraint satisfaction, and knowledge representation.
Week 11: Logic, probability, Bayesian networks, inference, and decision theory.
Week 12: Supervised and unsupervised machine learning.
Week 13: Reinforcement learning, model evaluation, and regularization.
Week 14: Neural networks, deep learning, CNNs, and transfer learning.
Week 15: Natural language processing, speech processing, and chatbots.