The University of Southampton
Courses

# COMP2208 Intelligent Systems

## Module Overview

This module aims to give a broad introduction to the rapidly-developing field of artificial intelligence, and to cover the mathematical techniques used by this module and by other artificial intelligence modules in the computer science programme

### Aims and Objectives

#### Module Aims

This module aims to give a broad introduction to the rapidly-developing field of artificial intelligence, and to cover the mathematical techniques used by this module and by other artificial intelligence modules in the computer science programme.

#### Learning Outcomes

##### Knowledge and Understanding

Having successfully completed this module, you will be able to demonstrate knowledge and understanding of:

• The principal achievements and shortcomings of AI
• The difficulty of distinguishing AI from advanced computer science in general
• The main techniques that have been used in AI, and their range of applicability
• Likely future developments in AI
• Basic differential and integral calculus
##### Subject Specific Practical Skills

Having successfully completed this module you will be able to:

• Select appropriately from a range of techniques when implementing intelligent systems
##### Subject Specific Intellectual and Research Skills

Having successfully completed this module you will be able to:

• Assess the claims of AI practitioners as they relate to `intelligence'
• Assess the validity of approaches to model intelligent processing
• Assess the applicability of AI techniques in novel domains

### Syllabus

Introduction to AI - Flavours of AI: strong and weak, neat and scruffy, symbolic and sub-symbolic, knowledge-based and data-driven. - The computational metaphor. What is computation? Church-Turing thesis. The Turing test. Searle's Chinese room argument. Calculus - Differentiation - standard rules; Newton's method for finding roots; partial differentiation; integration - standard integrals; integration by parts; numerical integration. Search - Finding satisfactory paths and outcomes; chosen from: depth-first and breadth-first, iterative deepening, evolutionary algorithms, hill-climbing and gradient descent, beam search and best-first. Finding optimal paths: branch and bound, dynamic programming, A*. Representing Knowledge - Production rules, monotonic and non-monotonic logics, semantic nets, frames and scripts, description logics. Reasoning and Control - Data-driven and goal-driven reasoning, AND/OR graphs, truth-maintenance systems, abduction and uncertainty. Reasoning under Uncertainty - Probabilities, conditional independence, causality, Bayesian networks, noisy-OR, d-separation, belief propagation. Machine Learning - Inductive and deductive learning, unsupervised and supervised learning, reinforcement learning, concept learning from examples, Quinlan's ID3, classification and regression trees, Bayesian methods. Key Application Areas, selected from: - Expert system, decision support systems - Speech and vision - Natural language processing - Information Retrieval - Semantic Web

### Learning and Teaching

TypeHours
Follow-up work18
Lecture36
Preparation for scheduled sessions18
Revision10
Total study time150

Pearl, J. (1988). Probabilistic Reasoning in Intelligent Systems.

Russell, S and Norvig, P (2003). Artificial Intelligence: A Modern Approach.

Mitchell, T. (1997). Machine Learning.

Nilsson NJ (1998). Artificial Intelligence: A New Synthesis.

Copeland J, (1993). Artificial Intelligence: A Philosophical Introduction.

### Assessment

#### Summative

MethodPercentage contribution
Coursework 50%
Exam  (2 hours) 50%

#### Referral

MethodPercentage contribution
Exam 100%

#### Repeat Information

Repeat type: Internal & External