Artificial Intelligence Education

The artificial-intelligence course published by Sharif University of Technology introduces intelligent agents through search, knowledge representation, inference, and machine learning. An agent is a system that selects actions in pursuit of an objective; this framing connects different techniques rather than treating artificial intelligence as a single application. The stated learning goals include developing solutions and understanding how methods work well enough to adapt them to particular problems. These are instructional objectives, not independently measured learning outcomes. s1

The computer-science curriculum documentation associated with Ferdowsi University of Mashhad provides another example. Its undergraduate AI outline moves from state-space search and constraint satisfaction to adversarial search, reinforcement learning, and Bayesian networks. These topics concern different problems: finding a route through possible decisions, satisfying rules, learning through rewards, or reasoning under uncertainty. Their inclusion demonstrates a broader academic scope than learning to operate a chatbot. s2

[s1] [s2]

Sharif’s published prerequisites include Python programming, data structures, and algorithms, with probability, statistics, and linear algebra strongly recommended. This combination links implementation with mathematical reasoning: students need to express a method in code and understand the assumptions behind it. The page does not present introductory AI as a substitute for these foundations. s1

The Mashhad outline instead explicitly lists foundations of logic and algorithm design and analysis as prerequisites. Its syllabus includes uninformed and informed search, the A-star search algorithm, game-tree reasoning, and reinforcement-learning methods. It also covers Bayesian-network representation and inference, including exact methods and approximation by sampling. These differences are useful evidence of how curriculum documents organize preparation; they should not be interpreted as a ranking of institutions or proof of uniform classroom delivery. s2

[s1] [s2]

A separate Sharif machine-learning course repository makes slides, Jupyter notebooks, and exercises available and directs readers to materials from earlier semesters. Jupyter notebooks combine executable code with explanation, making them a practical format for examining computational methods. The repository also points to Persian-language course videos and additional resources. Its public availability establishes access to teaching materials, rather than open admission to the university or the award of academic credit. s3

Taken together, these sources illustrate distinct layers of learning: foundational programming and mathematics, an introductory survey of intelligent methods, and more focused machine-learning study. That sequence is an editorial synthesis of the inspected materials, not a nationally mandated pathway. The sources explain what is taught more clearly than whom it reaches. A fuller account of AI education in Iran still requires evidence about institutions outside these examples, access to computing resources, participation, and the assessment of practical competence. s1s2s3

[s1] [s2] [s3]