Knowledge and Reasoning

Basic AI concepts (agents)
Basics of procedural programming

– 2 theoretical hours each week used to present new concepts and topics in the field of artificial intelligence focused on the representation of knowledge and reasoning using different intelligent algorithms. Presentation of subjects using slides and practical examples.


– 2 practical hours each week in which students have the opportunity to apply the concepts learned to solve concrete problems, using different tools and several practical cases.

Goals:


It is intended that students acquire knowledge in the area of intelligent systems, namely in the various forms of knowledge representation and reasoning methodologies, based on different algorithms in the area of artificial intelligence, namely rule-based systems, case-based systems, fuzzy systems, Bayesian networks, neural networks and clustering algorithms.


Skills
Knowledge and understanding:


• Explain the main forms of knowledge representation.


• Identify and explain the main types of intelligent reasoning
• Understand the main characteristics of different forms of knowledge and reasoning and forms of machine learning and be able to make accurate and well-founded choices to solve problems.

Application of knowledge:


• Acquisition and understanding of concepts essential to the representation of knowledge in intelligent systems.


• Acquisition and application of knowledge about understanding various forms of reasoning and forms of machine learning.


• Ability to design, implement, and evaluate intelligent applications using the various knowledge and reasoning methodologies discussed.

Communication:


• Prepare clear documentation as part of the development of practical work, identifying and justifying the main decisions made.

Autonomy and self-learning:

• Ability to carry out autonomous and group work.


• Development and autonomy in learning.
• Increase the ability to propose intelligent solutions to solve problems with different characteristics.

1) Forms of knowledge representation


2) Rule-based reasoning:


• Introduction to Expert Systems

• Generalities and Basic Principles

• Rules

• Inference Engine: Forward and Backward Chaining

3) Case-based reasoning:

• The RBC paradigm

• Aamodt & Plaza Diagram

• Case Representation and Memory Models

• Similarity Functions

4) Fuzzy Reasoning:

• Introduction to fuzzy logic

• Fuzzy sets and numbers

• Computation with linguistic terms

• Mamdani Inference

5) Probabilistic reasoning:

• Bayesian Networks

• Bayes theorem

• Joint and Conditioned Probability

• Construction of Bayesian Networks

• Application examples

6) Reasoning based on machine learning:

• 
Linear regression

• Neural Networks


              – Perceptron

              – Gradient Descent and delta rule

              – Multilayer networks

              – BackPropagation Algorithm

              – Other neural networks


• Clustering Algorithms


              – Density

              – Distribution

              – Centroid

              – Hierarchical

              – Comparison of various algorithms: K-means, DBSCAN, BIRCH, etc.

The final classification will be obtained based on the grades in the theoretical component and the practical component.


The theoretical component is scored for 12 points (in 20) and can be obtained through exam assessment.


The practical component is scored for 8 points (in 20) and consists of carrying out practical work.

To obtain approval for the curricular unit, a student must obtain a grade greater than or equal to 9.5 out of 20.

Assessment components:


• Theoretical:
The theoretical component grade is obtained by one written exam without consultation with a quotation of 12 points
• Practice:
The grade for the practical component corresponds to the grade obtained in the practical work to be carried out during the semester with a rating of 8 points.
The work has a mandatory defense on a date to schedule.

The work must be carried out in groups of two students.


The practical work can only be submitted once and is valid for the three exam periods.


The work is delivered through Moodle.


Students with access to the special/extraordinary season and students covered by other regimes (associative leader, high-competition athlete, etc.), who may require several dates to take the theoretical exam, must submit the practical work on the same date indicated above.

NAO