Overview

Large language models predict sequences of symbols. Many intelligent systems, however, must also represent states, actions, physical constraints and the consequences of intervention. World models address this broader problem.

This course introduces state representation, transition dynamics, latent models, planning, agents and model-based decision-making. It connects classical ideas from control and reinforcement learning to current work on learned environments and embodied systems.

The course is deliberately introductory. Its purpose is to show why world models are discussed as one possible direction beyond language-only systems and what technical problems remain unresolved.

Learning objectives

By the end of the course, learners should be able to:

• distinguish sequence prediction from modelling an environment with states and actions;
• explain transition models and latent representations;
• describe how world models can support planning and control;
• connect model-based reinforcement learning to modern learned simulators;
• identify challenges in partial observability, uncertainty and physical grounding;
• evaluate claims about world models with appropriate technical caution.

Intended participants and prerequisites

Course structure

The course is organised into six units. Each unit combines concise explanation, worked examples and short analytical exercises.

Unit	Subject	Principal questions
1	States, actions and environments	What must a model represent to reason about an interactive world?
2	Dynamics and prediction	How are transitions between states learned or specified?
3	Latent world models	Why model hidden representations rather than raw observations?
4	Planning and control	How can an internal model support decisions before acting?
5	Embodiment and partial observability	What changes when agents receive incomplete and noisy observations?
6	Beyond language-only systems	Which research problems motivate current interest in world models?

Teaching and assessment

Teaching is delivered through concise lectures, technical notes, worked examples and guided exercises. AI-supported interaction may be used to explain concepts, generate additional practice and challenge a learner's reasoning, but assessment remains based on the learner's demonstrated understanding.

Component	Purpose	Weight
Short exercises	Check understanding of the principal concepts	40%
Applied investigation	Use the course framework on a new problem or dataset	30%
Final assessment	Integrated reasoning across the course	30%

A certificate of completion is issued only when the required assessments have been completed at the stated standard. Participation alone does not constitute completion.

Faculty

Materials and technical requirements

Learning environment

Learners need a current web browser and access to the Gordon learning platform. Where a course uses code, spreadsheets, datasets or external readings, the relevant requirements are identified inside the unit.

Course materials

Required readings, datasets and practice materials are identified within each unit. Public courses are designed so that required core materials can be accessed without a separate tuition charge.