Skip to content

CS 204: Classical Artificial Intelligence
Course Code CS 204
Course Name Classical Artificial Intelligence
Department Computer Science
Semester Offered Even (Term 3 - Shanghai)
Tuition Hours 30 hours
Course Level Intermediate
Pre-requisite CS 105: Practical Machine Learning
Co-requisite -
Course Objective Modern AI is dominated by data-driven approaches. But not all problems are best solved with data. Some problems require structure, rules, reasoning, and explicit decision-making.

This course introduces the foundations of classical AI, where intelligence is built through search, logic, constraints, and knowledge representation rather than statistical learning.

Students will learn how to design systems that can plan, reason, and solve problems in well-defined environments, often with far greater efficiency and interpretability than black-box models.

By the end of the course, students will understand when to use rules instead of models, and how to combine both approaches to build robust AI systems.
Course Philosophy This course emphasizes
  • Reasoning before learning
  • Structure over brute force
  • Clarity over complexity
Students will learn that intelligence is not always about training bigger models. Sometimes it is about designing better systems. The goal is to develop the ability to choose the right approach for the problem.
Course Learning Outcomes Upon successful completion of this course, students will be able to:
  • Understand and implement search algorithms for problem-solving.
  • Model problems as state spaces and design efficient solutions.
  • Apply constraint satisfaction techniques to structured problems.
  • Represent knowledge formally using logical systems.
  • Design simple expert systems for rule-based decision making.
  • Understand planning algorithms and sequential decision processes.
  • Apply probabilistic reasoning in uncertain environments.
  • Identify when classical AI methods outperform machine learning approaches.
Course Author Sagar Udasi
MSc Statistics and Data Science with Computational Finance from The University of Edinburgh.
Contact: sagar.l.udasi@gmail.com
Course Organiser TBD
Details will be updated before course commencement.
No. Lecture Title Concepts Covered Lecture Objective
01 AI Before Neural Networks History, symbolic AI vs ML Build perspective on different approaches to intelligence.
02 What Does It Mean To Solve A Problem? Problem formulation, state spaces Teach structured thinking for problem-solving.
03 Searching Without Thinking Uninformed search, BFS, DFS Introduce brute-force approaches and their limits.
04 Searching With Intelligence Heuristic search, A* Improve efficiency using informed strategies.
05 When The Search Space Explodes Complexity, pruning Handle large problem spaces effectively.
06 Constraints Change Everything Constraint satisfaction problems Solve structured problems with constraints.
07 Making Decisions With Rules Rule-based systems, expert systems Build interpretable decision systems.
08 Representing Knowledge Explicitly Logic, propositional and predicate logic Teach formal reasoning about the world.
09 When Logic Meets Reality Inference, resolution Apply logical reasoning to real problems.
10 Planning Ahead Planning algorithms, goal-based systems Enable systems to act over time.
11 Sequential Decisions In The Real World Markov decision processes Model decision-making under uncertainty.
12 Probability As A Language Of Uncertainty Probabilistic reasoning basics Extend logic to uncertain environments.
13 Bayesian Thinking In AI Bayesian networks Combine probability with structure.
14 When Rules Beat Models Comparing symbolic AI and ML Teach when classical methods are better.
15 Hybrid Systems: Best Of Both Worlds Combining ML and symbolic AI Build robust, real-world AI systems.
16 Case Study: AI Without Data Real-world symbolic AI systems Show practical applications beyond ML.
17 Designing An Intelligent Agent Agent architectures Connect concepts to real AI systems.
18 From Logic To Action Decision pipelines Build end-to-end reasoning systems.
19 Applying Classical AI To Your Product Integration with student projects Directly connect to ongoing builds.
20 Demo Day: Can Your System Reason? Presentations, evaluation Students showcase reasoning-based systems.
Component Weightage
Problem Sets (Search & CSP) 30%
Rule-Based System Project 20%
Final Project: Intelligent Agent (Hybrid or Classical) 30%
Viva + Reasoning Evaluation 20%
Type Resource Provider
Lecture CS188 Introduction to AI UC Berkeley
Lecture MIT 6.034 Artificial Intelligence MIT OCW
Reading Artificial Intelligence: A Modern Approach Russell & Norvig
Reading Probabilistic Graphical Models Daphne Koller
Practice AI Search Visualizations Various online tools
Documentation NetworkX (Graph Algorithms) networkx.org