Monday, July 16

11am – 2pm

Registration and welcome lunch

2pm -3.30pm

Emilio Calvanese Strinati: 5G Networks and Beyond

In the last years, information communication, computation and storage technologies are jointly reshaping the way we use technology, meeting the future needs of a wide range of big data and artificial intelligence applications and, paving the way for a full customized autonomous user experience. In 2020 the 5G -Next Generation Communication Networks is expected to be operational and a global game changer from a technological, economic, societal and environmental perspective. 5G industry is intensively working today on designing, prototyping and testing fundamental technological advances to deliver the promised performance in terms of latency, energy efficiency, wireless broadband capacity, elasticity, etc. Nevertheless, many experts says that the next big step for cellular networks is not 5G, it is the cloud. This lecture will present the interdependencies between 5G KPIs, 5G key enabling technologies and the three levels of cloud: Fog, Mobile Edge Cloud and the Central Cloud. During the lecture it will be presented a vision on the technical enables of the beyond 5G networks and its further step to its cloudification. Special emphasis will be given to 5G and beyond open research challenges that researchers and industry are and will investigate in the near future.

3.30pm – 4pm

Coffee break

4pm – 5.30pm

Giuseppe Caire: Towards 5G heterogeneous wireless networks: massive MIMO, fog-massive MIMO, and mmWave small cells.

A key feature of 5G networks is the decoupling of data and control planes. This means that user devices can be controlled by some large cellular infrastructure offering ubiquitous coverage, while receiving/sending data at very high rate from/to any infrastructure device in their vicinity, in a clever and opportunistic fashion. In this talk I will review the basic principles and some recent results on three key access radio technologies that cover different aspects of the above mentioned heterogeneous 5G general vision. These are massive MIMO (providing coverage and access to a large number of devices), user-centric fog-massive MIMO (providing high data rates and very low delays for localized users), and mmWave small cells (providing very high data rates for very localized communication).

Tuesday, July 17

9am – 12.30pm (coffee break 10.30 am – 11am)

Giuseppe Caire: Recent Trends in Wireless Edge Caching.

Caching at the wireless edge has gained a lot of traction in the research community, although it is still at its infancy in terms of applications and products. This tutorial is an attempt to present the recent research in information theory, network coding, and wireless communications, related to caching. In particular, we shall depart from the traditional approach of “dynamic caching policy” and “hit rate maximization”, and focus on the stream of relatively recent papers that have characterized exact capacity and throughput scaling laws for an idealized scenario where the content library is fixed, the cache placement is performed a priori, and the goal consists o delivering the user requests in the shortest possible time. While this framework is clearly an over-simplification, it yields surprisingly clean and elegant optimality results for several network topologies and it is practically motivated by modern on-demand media delivery (streaming), where library and caches update and actual content transmission occur at very different time scales. A special emphasis shall be given to the interplay of edge-caching and physical layer schemes.

12.30pm – 2pm


2pm – 5.30pm (coffee break 3.30pm 4pm)

Giuseppe Bianchi: Network softwarization: towards data plane programmability

As defined in Wikipedia, Software Defined Networking (SDN) is about “decoupling the system that makes decisions about where traffic is sent (the control plane) from the underlying systems that forward traffic to the selected destination (the data plane)”. In the first part of the tutorial, we will overview the motivation that has lead, exactly a decade ago, to the emergence of SDN as well as the reasons and the pragmatic compromises (most notably in the OpenFlow Design choices) which have made it viable. We will then discuss recent evolution, deployment perspectives, relation with emerging network function virtualization technologies, as well as current limitations and open research issues, with specific focus to the programmability challenges of the data plane. In the second part of the talk, we will focus on a critical issue: how to meet performance and programmability at the same time. We will discuss device-level programming abstractions (e.g., P4, extended finite state machines) more expressive than OpenFlow, i.e., which permit to program and deploy inside the switch more complex and dynamic per-flow stateful “behavioral rules” rather than static forwarding rules. We will then provide some insights on high speed hardware architectures capable to support programmable network functions, discussing the relevant relation with network function virtualization and the emergence of NFV “accelerators” in the form of smart Network Interface Cards supporting such programming abstractions.

Wednesday, July 18

9am – 12.30pm (coffee break 10.30 am – 11am)

Sergio Barbarossa: Graph based signal analysis

Processing signals defined over discrete sets is a well established framework to extract information from data. Discrete-time signal processing or image processing are well known examples of signal processing over a discrete set, where the set is a one or a two-dimensional uniform grid. Generalizing this very old idea, we may associate signals to sets of points, or point cloud,  characterized by neighborhood relations that are much more general than regular grids.  In this talk, we introduce signal processing over simplicial complexes, as a very general way to incorporate multi-way relations among points. We start focusing on graph signal processing, where the graph is just a way to incorporate two-way relations, and then we generalize the approach to the general multi-way relation case. We introduce the basics tools to extract information from the data and to infer the graph or simplicial complex structure from the data. Then, we show applications to very diverse fields such as network data traffic, reconstruction of vehicular density from sparse samples, map of the electromagnetic field from sporadic data, inference of brain functional activity from electrocorticography (ECoG) data.

12.30pm – 2pm


2pm – 4.30pm STUDENTS PRESENTATIONS (coffee break 3.30pm 4pm)


Dinner at ristorante Bellavista Brunate

Meeting at 5.45 pm at ‘’Funicolare Como-Brunate’’ (See map)


Thursday, July 19

9am – 12.30pm (coffee break 10.30 am – 11am)

Alberto Testolin: Machine learning with artificial neural networks: history, state-of-the-art, future challenges

Recent theoretical and technical progress in artificial neural networks has significantly expanded the range of tasks that can be solved by machine intelligence. In particular, the advent of powerful parallel computing architectures based on graphic processing units, coupled with the availability of “big data”, now allows to create and train large-scale, hierarchical neural networks known as deep neural networks (LeCun, Bengio, & Hinton, 2015). These powerful learning systems achieve impressive performance in many challenging machine learning tasks, such as visual object recognition, speech processing, natural language understanding and automatic translation. In this seminar I will review the main theoretical foundations of artificial neural networks, discussing both supervised and unsupervised forms of deep learning, and sequential architectures based on recurrent networks. I will then provide examples and case studies related to a variety of cognitive tasks, as well as their applications on difficult optimization problems. During a practical, “hands-on session” students will be encouraged to train and analyze their own models on a simple pattern recognition problem.

12.30pm – 2pm


2pm – 5.30pm (coffee break 3.30pm 4pm)

Marco Dorigo: An Introduction to Swarm Intelligence

Swarm intelligence is the discipline that deals with natural and artificial systems composed of many individuals that coordinate using decentralized control and self-organization. In particular, the discipline focuses on the collective behaviors that result from the local interactions of individuals with each other and with their environment. Swarm intelligence research has two aspects: a scientific one, that studies existing natural systems; and an engineering one, that studies how to create, analyse, and control artificial swarm systems. Historically, scientific studies of swarm intelligence have inspired the design of artificial swarm systems. In this tutorial, after a short overview of some of the main natural swarm intelligence systems, I will focus on a number of different artificial swarm systems that have been proposed to solve difficult problems such as combinatorial and continuous optimisation problems, distributed and adaptive routing in packet switched networks, and the self-organised control of groups of autonomous robots.

Friday July 20

9am – 12.30pm (coffee break 10.30 am – 11am)

Mérouane Debbah: Random Matrix Theory for Wireless Communications

This tutorial introduces the field of Random Matrix Theory and particularly of Large Dimensional Random Matrix Theory from the point of view of their applications to Wireless Communications. The main objective is to provide the audience with the necessary tools and methods to handle wireless communication problems dealing with random matrices. A non-exhaustive list of important results is reviewed, the proofs of which are intended to provide the reader with both intuition and accurate tools in a step by step method.

12.30pm – 2pm


2pm – 3pm

Final Exam