AI and Expert Systems Lab Course: Exploration from Basic Theory to Laboratory Practice

This article introduces the CSE322 AI and Expert Systems Lab Course, which is aimed at 6th-semester students. By combining theory with laboratory practice, it helps students master core technologies such as basic AI algorithms (e.g., search, constraint satisfaction problems, game AI) and expert system construction (knowledge representation, inference engines, uncertainty handling). The course aims to cultivate students' practical skills, systematic thinking, and problem-solving abilities, laying a foundation for their learning and career development in the AI field.

The CSE322 course is guided by Professor Md. Mamun Hossain and targeted at 6th-semester students. The course objectives include: 1. Master core concepts of AI and expert systems; 2. Acquire the ability to implement basic AI algorithms; 3. Understand the overall architecture and design ideas of expert systems; 4. Apply AI technologies to solve practical problems; 5. Cultivate sensitivity to the development trends of AI technologies. This course builds a bridge from theoretical learning to practical application.

The basic AI experiment module covers three types of content:

1. Search Algorithms: Uninformed search (BFS, DFS, uniform cost, iterative deepening) and informed search (greedy, A*, IDA*, and heuristic function design);
2. Constraint Satisfaction Problems (CSP): Backtracking algorithms, constraint propagation (arc/node consistency), variable ordering (MRV/MD), value ordering (LCV);
3. Game AI: Minimax algorithm, Alpha-Beta pruning, evaluation function design, multi-player game extension.

The expert system construction experiment module includes:

1. Knowledge Representation: Logical representation (propositional/first-order predicate logic, reasoning, resolution), production rules (IF-THEN structure, forward/backward chaining, conflict resolution), frame representation (object frames, inheritance, default reasoning);
2. Inference Engines: Forward reasoning (pattern matching, working memory, inference cycle), backward reasoning (goal decomposition, backtracking, proof tree);
3. Uncertainty Handling: Probabilistic reasoning (Bayesian networks, exact/approximate reasoning), certainty factors (MYCIN system, CF propagation), fuzzy logic (membership functions, fuzzy rules, defuzzification);
4. Tools: Expert system shells such as CLIPS, Jess, Prolog, and experiments on knowledge acquisition and explanation facilities.

Learning Outcomes: Students can gain technical skills (programming, system design, debugging), cognitive abilities (abstract thinking, logical reasoning), and practical capabilities (experiment design, data analysis, document writing). Course Challenges and Solutions:

• Disconnect between theory and practice: Provide code examples and step-by-step guidance;
• System complexity: Modular design and step-by-step implementation;
• Computational resource constraints: Small-scale datasets, algorithm optimization, or cloud computing;
• Knowledge breadth: Focus on core concepts and classic algorithms, and provide extended reading materials.

Industry Applications: The skills cultivated in this course are applicable to professions such as AI engineers and data scientists, with application areas including fintech (credit assessment), healthcare (diagnostic assistance), intelligent manufacturing (predictive maintenance), etc. Future Trends: The course needs to integrate emerging technologies such as deep learning, reinforcement learning, and knowledge graphs; promote the evolution of expert systems towards interpretable, adaptive, distributed, and real-time reasoning directions to adapt to the development of AI technologies.