Avviso

Seminari su Statistical Relational Artificial Intelligence



Area:
Categoria:
Vari
Data avviso:
12-10-2018 
Descrizione breve:
Seminari su Statistical Relational Artificial Intelligence. 17 e 18 ottobre 2018 

 

Il 17 e 18 ottobre 2018 si svolgeranno due seminari su tematiche relative alla Statistical Relational Artificial Intelligence.
Il relatore sarà il Prof. Fabrizio Riguzzi dell'Università di Ferrara.

Agli studenti della Scuola di Economia che frequenteranno entrambi i seminari verranno riconosciuti 0,5 CFU lettera F.


mercoledì 17 ottobre 2018
ore 16-18  (Aula 20)

Title: Probabilistic Logic Languages.   SLIDES


Abstract: The combination of logic and probability is very useful for
modeling domains with complex and uncertain relationships among
entities. Many probabilistic logic languages have been proposed in
various research fields. In logic programming, the distribution
semantics has recently gained an increased attention and is adopted by
many languages such as the Independent Choice Logic, PRISM, Logic
Programs with Annotated Disjunctions and ProbLog. Other languages
instead follow a knowledge-based model construction approach in which
the probabilistic logic theory is used directly as a template for
generating an underlying complex graphical model. The talk will
illustrate these approaches for combining logic and probability and will
highlight similarity and differences. The talk will also introduce the
types of reasoning that can be performed with these languages:
inference, weight learning and structure learning. In inference we want
to compute the probability of a query given the model and, possibly,
some evidence. In weight learning we know the structural part of the
model (the logic formulas) but not the numeric part (the weights) and we
want to infer the weights from data. In structure learning we want to
infer both the structure and the weights of the model from data.

 



giovedì 18 ottobre 2018
ore 9-11  (Aula Informatica)

Title: Reasoning with Probabilistic Logic Programming Languages.   SLIDES

Abstract: The talk will survey existing approaches for inference and
learning in Probabilistic Logic Programming. It will discuss in details
algorithms for performing inference on Probabilistic Logic Programming
languages that follow the distribution semantics and in particular the
PITA algorithm that uses tabling and answer subsumption. The talk will
then concentrate on algorithms for learning models following the
distribution semantics, discussing first parameter learning and then
structure learning. Existing systems for parameter learning use either
gradient descent or the EM algorithm. The talk will present various
systems for parameter learning of probabilistic logic programs, focusing
especially on EMBLEM, that uses EM. Recently, structure learning systems
have started to appear, with promising initial results. Various search
strategies have been investigated in Probabilistic Inductive Logic
Programming. The system SLIPCOVER uses clause revision followed by
greedy theory search. All the presented systems are available in the web
application http://cplint.eu.


Speaker: Fabrizio Riguzzi

Fabrizio Riguzzi is Associate Professor of
Computer Science at the Department of Mathematics and Computer Science
of the University of Ferrara. He was previously Assistant Professor at
the same university. He got his Master and PhD in Computer Engineering
from the University of Bologna.

Fabrizio Riguzzi is vice-president of the Italian Association for
Artificial Intelligence and Editor in Chief of Intelligenza Artificiale,
the official journal of the Association.

He is the author of more than 150 peer reviewed papers in the areas of
Machine Learning, Inductive Logic Programming and Statistical Relational
Learning. His aim is to develop intelligent systems by combining in
novel ways techniques from artificial intelligence, logic and
statistics.




LINK UTILI. 

Scopri cosa vuol dire essere dell'Ud'A

SEDE DI CHIETI
Via dei Vestini,31
Centralino 0871.3551

SEDE DI PESCARA
Viale Pindaro,42
Centralino 085.45371

email: info@unich.it
PEC: ateneo@pec.unich.it
Partita IVA 01335970693

icona Facebook   icona Twitter

icona Youtube   icona Instagram