]]>

In this 6–hour lecture, we provide an introductory overview of Approximation Fixpoint Theory (AFT), an abstract, algebraic framework to capture semantics of logics based on fixpoints of a semantic operator.

We start by reviewing the main motivations that, 20 years ago, led to the conception of AFT, namely that several logics from non-monotonic reasoning (in particular logic programming, autoepistemic logic, and default logic) are defined on *the same* underlying principles. And that these principle can be captured by reasoning on an underlying semantic operator.

This leads us to the definitions of the main types of fixpoints studied in AFT and their different characterisations.

Next, we see some recent developments, such as extensions of the theory with a notion of groundedness and we focus on the broad variety of applications of AFT, ranging from abstract argumentation to database theory and to extensions of logic programming.

A strong focus lies on the answering the question *why does AFT work well for these applications?*.

Next, we see some recent developments, such as extensions of the theory with a notion of groundedness and we focus on the broad variety of applications of AFT, ranging from abstract argumentation to database theory and to extensions of logic programming.

A strong focus lies on the answering the question *why does AFT work well for these applications?*.

We conclude by discussing some potential topics for future work.

Bart Bogaerts studied a Master’s in mathematics at the KU Leuven, after which he started a PhD in Computer Science.

During this PhD and subsequent postdocs (at KU Leuven and Aalto University (Finlan)), he worked on various aspects

During this PhD and subsequent postdocs (at KU Leuven and Aalto University (Finlan)), he worked on various aspects

of knowledge representation and reasoning; an important recurring theme was semantic unifying frameworks, and in

particular Approximation Fixpoint Theory. Other topics of interest include representation languages and the

development of search algorithms, often with a focus on symmetry exploitation.

Since October 2018, he is an assistant professor at the Vrije Universiteit Brussel.

Monday 08.07.2019, 15.00-17:00 – Aula Seminari (Cubo 44Z)

Tuesday 09.07.2019, 15.00-17:00 – Aula Seminari (Cubo 44Z)

Friday 12.07.2019, 15.00-17:00 – Aula Seminari (Cubo 44Z)

]]>- 24.06.2019, ore 09:00 – 12:00, Distretto DOMUS residenze chiodo2 (vicino al cubo 41c)
- 26.06.2019, ore 09:00 – 12:00, Distretto DOMUS residenze chiodo2 (vicino al cubo 41c)
- 27.06.2019, ore 09:00 – 12:00, Distretto DOMUS residenze chiodo2 (vicino al cubo 41c)
- 28.06.2019, ore 09:00 – 12:00, Distretto DOMUS residenze chiodo2 (vicino al cubo 41c)

**PROGRAMMA**

**Ore 15.00** Apertura della Cerimonia

** Prof. Gino Mirocle Crisci*** *Magnifico Rettore dell’Università della Calabria

**Ore 15.15 ***Motivazioni del Conferimento*

** Prof. Luigi Palopoli **Direttore del DIMES dell’Università della Calabria

**Ore 15.30** *Laudatio del Prof. Carlo Zaniolo*

* ***Prof. Sergio Greco **Ordinario di Sistemi di Elaborazione delle Informazioni, Unical

**Ore 16.00 ***Lectio Magistralis: “*Declarative Languages and Algorithms for BigData Applications*”*

** Prof. CARLO ZANIOLO **Full Professor – University of California, Los Angeles (UCLA)

*Ore *17.00 *Conferimento del Dottorato di Ricerca Honoris Causa in Information and Communication Technologies*

*Lectio Magistralis: ***Declarative Languages and Algorithms for BigData Applications**

*Prof. Carlo Zaniolo*

The critical importance of advanced knowledge-based applications made possible by BigData underscores the need for high-level declarative languages providing ease-of-use, portability, scalability and performance on such applications. Researchers have pursued this ambitious goal by extending the enabling technologies of Relational Databases and Logic Programming, making great progress that benefited their fields and commercial systems. Yet, progress has been hampered by non-monotonic reasoning issues that have limited our ability to express complex algorithms using negation and aggregates in recursive queries. Major progress on this front has recently been achieved with the introduction of the concept of Pre-Mappability (PreM) that makes possible to view recursive programs with aggregates as stratified programs—stratification is a simple syntactic notion that assures declarative semantics and portability. PreM is a very general notion that applies to diverse constraints and languages and also dove-tails with map-reduce and data streams. In fact, Apache-Spark SQL extended with PreM outperforms on graph applications special-purpose database systems created to exploit this area of weakness of commercial SQL systems. In our seminar, we will finally discuss how easily classical algorithms can be expressed concisely in Datalog with aggregates, and how PreM in these algorithms can be proven using the simple methods we discovered.

**Prof. Carlo Zaniolo: short biography**

Carlo Zaniolo was born in Vicenza, Italy. He received an E.E. Engineer degree at Padua University in 1968, and M.S. and Ph.D. degrees in Computer Science at UCLA in 1970 and 1976, respectively. After working at Bell Laboratories, Murray Hill, NJ, and MCC in Austin Texas, Prof. Zaniolo joined the UCLA CS Department in 1991 as a Full Professor of Computer Science, and was awarded the N.E. Friedmann Chair in Knowledge Science. At UCLA, he is now the co-director of the Scalable Analytics Institute. Prof. Zaniolo’s interests include big data and knowledge based systems, non-monotonic and temporal reasoning, internet information systems, answering questions, queries and searches in knowledge base. Prof. Zaniolo has published more than 300 papers in different areas but his fame is primarily due to his contributions to Database technology: his discovery of Multivalued Dependencies that he introduced in his PhD thesis; his work on null values, and on algorithms for relational schema design including the simplified definition of Third Normal Form(3NF) used in all textbooks; his work on a model called GEM, which extends the relational model with object-oriented features; his seminal contributions on Datalog and non-monotonic reasoning focused on stable models and their connection to choice models; his work on supporting and optimizing regular expressions in query languages. A most enduring theme of his esearch has been taming the non-monotonic nature of aggregate functions to let them serve as the basis for powerful queries, declarative algorithms, and knowledge-based applications (including declarative BigData applications).

]]>**Instructor:** Prof. Stefano Basagni (Northeastern University, USA)

**Abstract.:**

This course is intended to provide introduction to the latest research directions on methodologies and technologies for the emerging paradigm of

the Internet of Things (IoT). In particular, we will explore networking challenges and solutions for a compelling scenario where key IoT applications are

being defined and developed, namely, for underwater wireless networking. Particularly, the course will provide details on the following:

1) Motivations:

Applications, relevant funding programs, trends and tendencies.

2) Challenges and barriers to progress, for the IoT in general, and for the Internet of

Underwater Things (IoUT) in particular.

3) State-of-the-Art: Current research on underwater wireless communication and networking.

4) Tools for underwater networking protocol design. Multi-modality; energy harvesting, and learned routing for the variable channels.

5) Research directions and open problems.

**Date:**

- 13.05.2019, ore 10:00-13:00 – Aula Seminari DIMES (Cubo 42C – V Piano)
- 14.05.2019, ore 10:00-13:00 – Aula Seminari DIMES (Cubo 42C – V Piano)
- 15.05.2019, ore 10:00-13:00 – Aula Seminari DIMES (Cubo 42C – V Piano)
- 16.05.2019, ore 10:00-13:00 – Aula Seminari DIMES (Cubo 42C – V Piano)

Prof. Anatoly Zhigljvsky, Cardiff University, United Kingdom

The main part of the course will be devoted to the exposition of the classical theory of experimental design in regression including equivalence theorems and numerical methods for construction of optimal designs. At the second part, several recent advances in the theory of optimal design for correlated observations will be discussed.

**Contents: **

1. Introduction: general scheme of a scientific experiment, statistics versus design; main principles for designing experiments (simplicity and efficiency), introductory examples

2. Several basic schemes for construction of factorial designs: 2^(m-k) designs, Latin designs, BIBD.

3. Brief review of basic facts from regression analysis. Definition of design efficiency and optimality.

4. Optimality (equivalence) theorems in optimal designs, examples.

5. Numerical methods for construction of optimal designs.

6. Design for correlated observations: best linear unbiased estimation (BLUE) and corresponding optimal designs.

7. Some other areas where designing of experiments has proved to be very useful.

**Orari:**

Lunedì, 8 aprile 2019, 10.30-12.30 – Aula Seminari DIMES

Martedì, 9 aprile 2019, 9.30-11.30 – Aula Seminari DIMES

Mercoledì, 10 aprile 2019, 16.30-18.30 – Aula Seminari DIMES

Nei giorni 10-11-12 Aprile 2019 dalle ore 10:00 alle ore 12:00 nell’ Aula Seminari del DIMES – V piano, cubo 42/C

Per informazioni: Prof. Sandra Costanzo, costanzo@dimes.unical.it

Course Content

**Day 1**

1. Forward Models of Microwave Imaging

1.1. Field-based Integral Solutions of the Scattering Problem in Time and Frequency

1.2. Born and Rytov Approximations of the Forward Model of Scattering

1.3. Scattering Parameters and Integral Solutions in terms of S-parameters

1.4. 2D Model of Tomography in Microwave Scattering

**Day 2**

2. Linear Inversion Methods

2.1. Image Reconstruction of Pulsed-radar Data

a. Synthetic Focusing: Delay and Sum Method

b. Deconvolution Methods

2.2. Image Reconstruction of Frequency-domain Data

a. Qualitative and Quantitative Imaging with Microwave Holography

b. Qualitative Imaging with Sensitivity Maps

c. Quantitative Imaging with Scattered Power Mapping

**Day 3**

3. Performance Metrics in Imaging

3.1. Spatial Resolution

3.2. Dynamic Range

3.3. Data Signal-to-noise Ratio

4. Overview of Nonlinear Inversion Methods

4.1. Direct Iterative Methods

4.2. Model-based Optimization Methods

5. Microwaves in Tissue Imaging – Challenges and Advancements

**Abstract:**

The last decade has witnessed dramatic decrease in the price and size of microwave electronics and the advent of the radio-on-a-chip and the single-chip radars. This spurred unprecedented growth in microwave imaging and sensing, a technology which allows for “seeing” through optically opaque barriers and which is at the frontier of wireless research and development. In this course, we will introduce the methods of real-time microwave imaging. We will discuss hardware as well as reconstruction algorithms. The most common reconstruction algorithms such synthetic focusing, microwave holography, and scattered-power mapping will be discussed and illustrated through examples. Advanced optimization-based imaging algorithms will be also presented briefly. In conclusion, we will review the challenges and the advances faced by the application of microwaves in tissue imaging for cancer diagnostics.

**Short Biography of Instructor:**

Natalia K. Nikolova (IEEE S’93–M’97–SM’05–F’11) received the Dipl. Eng. (Radioelectronics) degree from the Technical University of Varna, Bulgaria, in 1989, and the Ph.D. degree from the University of Electro-Communications, Tokyo, Japan, in 1997. From 1998 to 1999, she held a Postdoctoral Fellowship of the Natural Sciences and Engineering Research Council of Canada (NSERC) at two Canadian universities, Dalhousie University in Halifax and McMaster University in Hamilton. In 1999, she joined the Department of Electrical and Computer Engineering at McMaster University, where she is currently a Professor. Her research interests include inverse scattering, microwave imaging, as well as computer-aided analysis and design of high-frequency structures and antennas. Prof. Nikolova has authored more than 260 refereed manuscripts, 5 book chapters, and the book “Introduction to Microwave Imaging” published by Cambridge University Press in 2017. She has delivered over 40 invited lectures around the world on the subjects of microwave imaging and computer-aided electromagnetic analysis and design. Prof. Nikolova is a Canada Research Chair in High-frequency Electromagnetics. She is a Fellow of the IEEE and a Fellow of the Canadian Academy of Engineering (CAE). She served as an IEEE Distinguished Microwave Lecturer from 2010 to 2013.

]]>

**Date e orari:**

26 Marzo 8:30-11:30 – Aula Seminari DIMES

27 Marzo 8:30-11:30 – Aula Seminari DIMES

28 Marzo 8:30-11:30 – Aula Seminari DIMES

29 Marzo 8:30-11:30 – Aula Seminari DIMES

**Abstract:**

The main aim of the course is the introduction of the main techniques based on the ensemble learning paradigma for the classification of large datasets (i.e., Boosting, Random Forest, Stacking, etc.).

The course will also illustrate some innovative techniques, based on recent research trends in the fields of ensemble, such as specialized ensemble, unsupervised ensemble and meta-ensemble.

In addition, real cases in the cybersecurity domain will be discussed together with some possible solutions based on ensemble techniques. Practical sessions will be devoted to the development of ensemble-based algorithms by using the Scikit-Learn development tool and Python.

**Programma:**

Introduzione agli ensemble.

I weak classifier. Selezione e fusione. Perché funzionano. Il voting. Architettura di un ensemble. Generalizzazione e diversità. Ensemble trainable enon trainable.

I metodi principali per la costruzione di ensemble.

Bagging. Boosting. Adaboost.M1 e Adaboost.M2. Random Forest. Stacking.

Ensemble: argomenti avanzati.

I meta-ensemble. Gli ensemble specializzati. Gli ensemble non supervisionati. Ensemble e Big Data.

Librerie per lo sviluppo di algoritmi basati sugli ensemble: scikit-learn e Python Data Analysis Library.

Un caso studio: la cybersecurity.

Meta-Ensemble per la classificazione di utenti. Ensemble specializzati per la classificazione delle intrusioni.