Artificial Intelligence

RICCARDO ROSATI
RICCARDO ROSATI

General objectives:

To know the main languages of the current semantic technologies, in particular, the families of class-based and rule-based knowledge representaton formalisms, and the main reasoning techniques for such formalisms. To know the standard semantic technologies based on the above knowledge representation formalisms, in particular the RDF language and the OWL language, with the goal of designing and managing an ontological knowledge base. To know the basic elements of the representation of actions and reasoning about actions.

Specific objectives:

Knowledge and understanding:
Description Logics (the main class-based knoeledge representation formalisms) and the main rule-based languages, in particular Datalog and some of its extensions. The main Web standards for semantic technologies, in particular the RDF, SPARQL and OWL languages.

Applying knowledge and understanding:
To be able to design a knowledge base, choosing the most appropriate formalism and technologies for the given application context.

Making judgements:
To be able to evaluate the main semantic aspects of a knowledge base and of a knowledge-based application. To be able to choose the best available technology for processing a knowledge base.

Communication skills:
The practical activities and the exercises allow the student to be able to communicate and share the requirements of an application requiring the construction and management of a knowledge base and/or the usage of the standard semantic technologies.

Learning skills:
Besides the classical learning skills provided by the theoretical study of the teaching materials, the practical activities stimulate the student to autonomously deepen her/his knowledge about some of the course topics, to teamwork, and to the practical application of the notions and techniques learned during the course.

- Knowledge of the main approaches to knowledge representation in Artificial Intelligence.
- Knowledge of some of the main automatic reasoning techniques in knowledge representation.
- Ability to apply previous approaches and techniques to specific use cases.

No prerequisites.

1 - Introduction to knowledge representation
2 - Class-based formalisms
Description Logics
Reasoning in Description Logics
Description Logics vs. relational databases
3 - Rule-based formalisms
Brief introduction to logic programming
Datalog
Reasoning in Datalog
Datalog vs. Description Logics
Datalog extensions
Datalog with negation
Answer Set Programming (ASP)
Reasoning in ASP
Comparison with SQL
4 - Semantic technologies
Semantic Web
RDF, RDFS, SPARQL
Linked data
Ontologies
OWL
OWL profiles
Reasoning in OWL profiles
5 - Knowledge representation and Deep Learning
Deep Learning and Large Language Models
Knowledge graph embedding
The role of Knowledge Representation and Reasoning in Large Language Models

Lecture notes distributed by the teacher.

Traditional face-to-face lectures.

No attendance obligation.

The exam consists of a written test and a practical project.

Examples and exercises are available on the Web and can be accessed starting from the page https://www.diag.uniroma1.it/rosati/krst/

• Knowledge Representation in AI


• Logic-based knowledge representation


• Class-based knowledge representation and reasoning


• Rule-based knowledge representation and reasoning


• RDF, SPARQL, OWL


• Knowledge representation in Machine Learning