Methods for dealing with uncertainty and imprecision in knowledge representation and revisionA project funded byIstituto Nazionale di Alta Matematica
"F. SeverI" - Gruppo
Nazionale di Calcolo Scientifico (INdAM- GNCS)
|
The project in brief
Uncertainty and
imprecision are pervasive in several real world domains. In
some cases, uncertainty is inherent to data collection and
processing. In other cases, imprecision might be tolerated in
order to augment the interpretability of knowledge. For
instance, in the so-called Semantic Web, several forms of
uncertainty can be found. They are due to the intrinsic
incompleteness of knowledge bases, arising from the
large-scale application context (the Web), but also from the
inconsistencies introduced by heterogeneous and distributed
interconnected data sources (the so-called Linked Data).
Uncertainty and imprecision need to be properly dealt with by AI-based systems. The most commonly used techniques are founded on probabilistic logic and fuzzy logic. These logics however are deeply different from the semantic viewpoint. In particular, the former is good for modeling uncertainty due to the randomness of phenomena. Conversely, the latter is more appropriate for representing imprecise information, namely information which can be applied gradually to world entities, which is typical to perception.
This project aims at investigating several methods for dealing with uncertainty and imprecision in knowledge representation and revision. In particular, it relies on formal tools such as probabilistic logic and fuzzy logic to formulate innovative hybrid solutions, with particular reference to the application context of the Semantic Web.
Uncertainty and imprecision need to be properly dealt with by AI-based systems. The most commonly used techniques are founded on probabilistic logic and fuzzy logic. These logics however are deeply different from the semantic viewpoint. In particular, the former is good for modeling uncertainty due to the randomness of phenomena. Conversely, the latter is more appropriate for representing imprecise information, namely information which can be applied gradually to world entities, which is typical to perception.
This project aims at investigating several methods for dealing with uncertainty and imprecision in knowledge representation and revision. In particular, it relies on formal tools such as probabilistic logic and fuzzy logic to formulate innovative hybrid solutions, with particular reference to the application context of the Semantic Web.
Project team
Coordinator:
Francesca Alessandra LISI
Department of Computer Science, University of Bari “Aldo Moro”, Italy
Unità INdAM Bari
Department of Computer Science, University of Bari “Aldo Moro”, Italy
Unità INdAM Bari
Participants:
- University of Bari (Francesca A. Lisi, Corrado Mencar)
- University of Calabria (Mario Alviano, Francesco Ricca)
- University
of Ferrara (Marco Alberti, Elena
Bellodi, Giuseppe Cota, Marco
Gavanelli, Evelina Lamma, Fabrizio
Riguzzi, Riccardo Zese)
- University
of Palermo (Roberto Pirrone)
Workshop
(Bari, 16 December, 2019)
Dipartimento di
Informatica, Università degli Studi di Bari "Aldo
Moro"
II piano, Aula 2A
II piano, Aula 2A
Schedule
| 09:00-09:15 |
OPENING |
| 09:15-10:30 |
SESSION
1 |
| Thomas
Lukasiewicz (University of Oxford, UK): Recent
Advances in Querying Probabilistic Knowledge
Bases |
|
| Francesca Alessandra LISI
(University of Bari): Learning and Reasoning
with Fuzzy Ontologies: Summary of research
from Bari |
|
| 10:30-11:00 |
COFFEE
BREAK |
| 11:00-12:30 |
SESSION
2 |
| Fabrizio Riguzzi, Evelina Lamma (University of Ferrara): Probabilistic reasoning and learning at UNIFE | |
| Fabrizio Riguzzi, Elena
Bellodi, Riccardo Zese, Marco Alberti, Evelina
Lamma (University of Ferrara): Probabilistic
Inductive Constraint Logic |
|
| Marco
Gavanelli, Evelina Lamma, Riccardo Zese,
Fabrizio Riguzzi, Elena Bellodi, Marco Alberti (University of Ferrara):
Probabilistic SCIFF |
|
| Marco
Alberti, Evelina Lamma, Fabrizio Riguzzi,
Riccardo Zese (University of Ferrara): Probabilistic
Hybrid Knowledge Bases |
|
| 12:30-14:00 |
LUNCH
BREAK |
| 14:00-15:30 |
SESSION
3 |
| Roberto
Pirrone (University of Palermo): Tensor
Conceptual Spaces |
|
| Corrado
Mencar (University of Bari): Granular
counting of uncertain data |
|
| Francesco
Ricca (University of Calabria): Beyond NP:
Quantifying over Answer Sets |
|
| Mario
Alviano (University of Calabria): Fuzzy
Answer Set Programming via Satisfiability
Modulo Theories |
|
| 15:30-16:00 |
COFFEE
BREAK |
| 16:00 - 17:00 | DISCUSSION & CLOSING |
 |
 |