Undergraduate Course: Informatics 2D - Reasoning and Agents (INFR08010)
|School||School of Informatics
||College||College of Science and Engineering
|Credit level (Normal year taken)||SCQF Level 8 (Year 2 Undergraduate)
||Availability||Available to all students
|Summary||This course focuses on approaches relating to representation, reasoning and planning for solving real world inference. The course illustrates the importance of (i) using a smart representation of knowledge such that it is conducive to efficient reasoning, and (ii) the need for exploiting task constraints for intelligent search and planning. The notion of representing action, space and time is formalized in the context of agents capable of sensing the environment and taking actions that affect the current state. There is also a strong emphasis on the ability to deal with uncertain data in real world scenarios and hence, the planning and reasoning methods are extended to include inference in probabilistic domains.
1. Intelligent Agents: Introduction
* Nature of agents, performance measures and environments
* Wumpus World Problem : An example thread (Programming environment) setup
2. Search based Planning
* Planning as a Search Problem: In deterministic, observable, static and known environments
* Smart Searching 1: Using constraints
* Smart Searching 2: Exploiting subproblems/Memoisation
* Informed Search and Exploration for agents
3. Logical Representation and Planning
* Propositional Logic Revisited (Shortcomings)
* First Order Logic & Encoding facts/rules in FOL
* Inference Rules for Propositional & FOL Calculus
* Unification and Generalized Modus Ponens
* Resolution based Inference and directing search with it
* Knowledge representation : Using FOL to represent action, space, time -- Wumpus Example
* Situation Calculus: Representing time in plans
4. Scaling Planning for Complex Tasks
* Representing States, Goals and Actions in STRIPS
* Partial Order Planning
* Planning and Acting in the Real World
5. Acting in Uncertain (real world) Environments
* Representation with Bayes Net
* Probabilistic Reasoning in Bayes Net
* Planning under Uncertainity : Wumpus world revisited
* Probabilistic Reasoning over Time I: hidden markov models
* Probabilistic Reasoning over Time II: dynamic Bayesian networks
* Markov Decision Processes
Relevant QAA Computing Curriculum Sections: Artificial Intelligence, Human-Computer Interaction (HCI), Intelligent Information Systems Technologies, Simulation and Modelling
Information for Visiting Students
|High Demand Course?
Course Delivery Information
|Academic year 2022/23, Available to all students (SV1)
|Learning and Teaching activities (Further Info)
Lecture Hours 30,
Seminar/Tutorial Hours 10,
Supervised Practical/Workshop/Studio Hours 8,
Summative Assessment Hours 2,
Programme Level Learning and Teaching Hours 4,
Directed Learning and Independent Learning Hours
|Assessment (Further Info)
|Additional Information (Assessment)
||The coursework component, worth 30% of the overall grade of the course, will consist of two assignments, worth 15% each.
||Hours & Minutes
|Main Exam Diet S2 (April/May)||2:00|
|Resit Exam Diet (August)||2:00|
On completion of this course, the student will be able to:
- Use task constraints to make search efficient.
- Perform Inference with First Order Logic and appreciate the strengths and weaknesses of this and other logic representations (eg Propositional).
- Use PDDL to plan and execute actions using either Propositional or First Order Logic representations.
- Create and reason with a representation of a Bayesian agent for handling a non-deterministic planning problem.
- Constructively engage in both self-study and peer-learning.
|* Russell, S. & Norvig, P., "AI: A Modern Approach", Prentice Hall or Pearson, 2003. 2nd Edition.|
* Thompson, S., "Haskell: The Craft of Functional Programming", Addison Wesley, 1999.
|Course organiser||Prof Alex Lascarides
Tel: (0131 6)50 4428
|Course secretary||Miss Kerry Fernie
Tel: (0131 6)50 5194