# Interdisciplinary Seminar

## Table of Contents

## 1 Home

Welcome on the webpage of the Interdisciplinary Seminar of the Doctoral School in Science and Engineering and the Doctoral Programme in Mathematics and Applications.

The goal of the Interdisciplinary Seminar is to provide an interdisciplinary environment where researchers can present and discuss cutting-edge research. the tasks will be aimed at a broad audience, with more emphasis on conceptual than technical issues. The research presented will vary, ranging from new methodological developments to applied empirical papers that use methodology in an innovative way.

## 2 Seminar details

- Title
- Interdisciplinary Seminar
- Category
- 2: Inter/cross-disciplinary competences, common academic and scientific modules
- Mandatory
- no
- Teacher
- -
- Certifier
- Simon Campese
- Responsible
- Simon Campese
- Number of ECTS
- 1
- Workload
- 25h presence and self-study
- Frequency
- Every summer term
- Multiple validation
- no
- Language
- English
- Maximum participants
- 20
- Admission criteria
- Validation will only be given one time during the whole PhD thesis. This is why priority will be mostly given to third year candidates, after a selection based on a short proposal.
- Evaluation/validation criteria
- To validate the module, the PhD candidate will be asked to:
- deliver a short presentation
- write a report on a talk in a domain different from his/her area of expertise
- participate regularly and actively
- help with the organisation of the seminar (by being the chair of or organizing food for a session)

- Learning outcomes
- introduce the interdisciplinary approach
- give the opportunity to discuss mathematical tools used in different scientific disciplines from an interdisciplinary point of view
- open to other subjects or different points of view on a given subject

- Description
- The goal of the Interdisciplinary Seminar is to provide an interdisciplinary environment where researchers can present and discuss cutting-edge research. the tasks will be aimed at a broad audience, with more emphasis on conceptual than technical issues. The research presented will vary, ranging from new methodological developments to applied empirical papers that use methodology in an innovative way.

## 3 Sessions

In each session, two talks will be given. Due to the current COVID-19 situation, at least the first sessions take place virtually via webex on Thursdays from 12:15-13:15. Scheduled sessions so far are (see below for titles, domains, abstracts and a link to the webex meetings):

date | speaker 1 | speaker 2 | chair |
---|---|---|---|

Sergio Helgueta | Ahmed Soliman | Alisa Govzman | |

Petra Sala | Nikola de Lange | Gregoire Defoort | |

Ludovica Paseri | Xavier Mazur | Xavier Mazur | |

Alisa Govzman | Gregoire Defoort | Petra Sala |

### 3.1 upcoming sessions

#### 3.1.1 (12:15-13:15)

- talk 1
**Homotopy in pictures**(Alisa GOVZMANN)

Domain: Mathematics (Topology)

Abstract: Homotopy was first discovered in the context of complex integration by Cauchy some 200 years ago and then extended by Riemann and Jordan to the context of classifying some low dimensional topological spaces. Nowadays Homotopy Theory gives rise to a huge branch of modern mathematics, where one of the main goals is still the classification of topological spaces. Even though this theory can get quite complicated and abstract, in my talk, I will describe the basic idea behind the concept of Homotopy just using pictures and animations. Not a single formula will be used and yet, by the end of the talk you will know some interesting topological spaces and will be able to classify them up to Homotopy.

- talk 1
**From models to experiments, a brief history of the atom**(Grégoire Defoort)

Domain: Material science

Abstract: The way particles are arranged follow strict rules: from solid to gas, numerous models intend to picture the physical and chemical “dance” that binds atoms, particles and molecules together. From the very first model described by the Greek philosopher, (atoms, or what is indivisible), to now-a-days complex quantum equations, we have come far, yet, results obtained in laboratory with the most precise machines shows some discrepancies between models and experimentations. With such level of entanglement and complexity, simulations are often used to understand the origins of these discrepancies and provide a practical solution that can be translated in laboratory experiments. During this presentation, we will have a brief explanation on what are atomistic models and how are them used to describe matter. Furthermore, we will try to understand what are the tools used in this domain and how can simulations (jointly with experimentations) are used to have a better understanding of solid-state physic.

### 3.2 past sessions

#### 3.2.1 (12:15-13:15)

chair: Xavier Mazur

link to webex meeting (password: BczswZEP733)

- talk 1
**European Open Science Cloud: when cloud computing and HPC enhance scientific research.**(Ludovica Paseri)

Domain: Law and Technology, Data Protection, Open Science

Abstract: The European Union in recent years has expressed a clear willingness to foster the development of cloud computing in Europe as much as possible. One of the first initiatives is to help researchers to spread and share their knowledge: this can happen thanks to the European Open Science Cloud, a trusted and federated environment connecting researchers across Europe, combining Open Science Projects, cloud technologies and HPC platforms. How is this European cloud initiative articulated? Which is the link between the European strategy and the Open Science projects? What are the legal challenges that this ambitious project raises?

- talk 2
**Controllability of transportation networks : Impact of controllers**(Xavier Mazur)

Domain: Transport engineering, Complex network, Optimization

Abstract: With a rising number of road users, transportation networks became more and more subject to negative effects like congestion. In order to maximize their efficiency and avoid delays for road users, advanced control strategies are developed to use the existing infrastructures at their full potential. These control strategies employ controllers, such as traffic light or pricing toll, to redirect road users toward less used routes, to have a more equitable use of the infrastructure and avoid the congestion of the network. However, little attention has been devoted to investigating how different locations, kinds and amounts of controllers in a network affect the overall performance of network-wide schemes.

#### 3.2.2 (12:15-13:15)

chair: Grégoire DEFOORT

link to webex meeting (password: EWbkZ4spP73)

**Crypography: a stroll along the PAKE protocol**(Petra SALA)

Domain: Computer Science, Cryptography, Security Protocols

Cryptography is a large area in Computer Science that makes sure that information and communication remains safe and makes it as difficult as possible for any adversary to try to get a hold of private information, such as usernames and passwords and prevents the adversary of misusing them. Some of the interesting objects in Cryptography that are used in daily life are Digital Signatures, Zero Knowledge proofs, TLS, Encryption, Security protocols, etc. I will be talking about PAKE (Password authenticated key exchange protocol) protocols whose goal is to establish a high entropy shared secret key between two users who share a password and communicate over a hostile network. These protocols are based on Diffie- Hellman key exchange and there are different models which provide security proofs, making PAKE desirable tool to use for key exchange. I will be also talking about abilities that the adversary has when attacking the protocol, along with different attacks he can perform.

**A short bioinformatics workflow**(Nikola DE LANGE)

Domain: Biology, Bioinformatics

Processing and analysis of high-dimensional, complex biological data requires a variety of heterogenous software. Given the large volume of data produced, high performance computing is necessary to perform such analyses in time. Bioinformatics tools are developed individually and differ in file format as well as CPU and memory requirements. This frequently leads to individual scripts for each step in a pipeline and manual intervention for data management. Long scripts automating all steps often reserve vastly more computing resources than necessary. Workflow management system such as Snakemake have become invaluable for fast and resource-efficient computational analyses. In this talk we will have a look at an example bioinformatics workflow and see what improvements can be achieved with a workflow management system.

#### 3.2.3 (12:15-13:15)

chair: Alisa Govzmann

link to webex meeting (password: MpVv7mZ5Aj3)

**Epigenetics: Meaning and Applications in Scientific Research, focusing in the context of Parkinson’s Disease**(Sergio Helgueta Romero)

Domain: Biology,Neurogenetics

Abstract: Epigenetics can be defined in multiple ways. One explanation could be: “The way in which the products from the genes can alter the activation of the first ones, changing the Phenotype but no the Genotype”. Genotype is “the set of genes of an organism”, while Phenotype is “all of observable characteristics of an individual resulting from the interaction of its genes with the environment”. In this talk we will travel around these concepts by the hand of Mendel, Lamarck and Darwin, who will help us to understand them. During this trip, we will review how the genes, that are inherited from our parents (Genetics), are “expressed” producing proteins, and how the proteins, which carry out different functions, can also modify the activation of the genes (Epigenetics) according to the internal environmental signals. Concepts you already know such as DNA, RNA, genes or proteins; and others you likely never heard about like transcription factors, gene expression or enhancers will be introduced as well. Finally, we will discover why studying Epigenetics in scientific research is important, especially in the context of diseases, focusing in the second most common neurodegenerative disease: Parkinson’s Disease.

**Introduction to Biomechanics: from motion analysis to developing implants**(Ahmed Soliman)

Domain: Biomechanics

Abstract: Biomechanics is the science of studying the mechanical aspects of musculoskeletal structure, function, and motion during the movement of a living body, including how muscles, bones, tendons, and ligaments work together to produce movement. The study of biomechanics is important when determining what causes injuries, how to prevent them re-occurring, and how to treat it invasively by using implants or noninvasively following rehabilitation procedures in order to restore the musculoskeletal system functionality. The outcomes of a recently completed study by the University of Luxembourg and the Centre Hospitalier de Luxembourg will be presented. The stability provided by two pelvis implants for pelvic ring fractures was investigated numerically and experimentally with a favourable recommendation for the orthopedic procedure developed and adopted by the Centre Hospitalier de Luxembourg.

## 4 Archive

### 4.1 seminar 2019:

date | room | speaker 1 | speaker 2 | chair | food |
---|---|---|---|---|---|

1.040 | - [welcome meeting] | - | - | Simon Campese | |

1.040 | Jeroen van Wier | Diptaishik Choudhury | Antonio di Maio | Eduardo Ibarguengoytia | |

1.040 | Alessia Sciortino | Tieu Long Mai | Abdallah Ibrahim | Li Chen | |

1.050 | Massimo Notarnicola | Itzel Vazquez Sandoval | Alexey Kalugin | Sebastiano Tronto | |

1.050 | Abdallah Ibrahim | Balasz Pejo | Axel Chemla | Massimo Notarnicola | |

1.050 | Sebastiano Tronto | Céline Barlier | Tieu Long Mai | Sina Dortaj | |

1.050 | Pietro Sgobba | Tijana Randic | Jeroen van Wier | Alessia Sciortino | |

1.050 | Eduardo Ibarguengoytia | Axel Chemla | Itzel Vazquez Sandoval | Diptaishik Choudhury | |

1.050 | Alexey Kalugin | Najmeh Soroush | Balasz Pejo | Tijana Randic | |

1.050 | Sina Dortaj | Li Chen | Céline Barlier | Pietro Sgobba |

#### 4.1.1 (12:30pm-13:30pm, room MNO 1.050)

chair: Itzel Vazquez Sandoval

food: Diptaishik Choudhury

- talk 1
**A simple approach to Quantum Mechanics**(Eduardo Ibarguengoytia)

Domain: Mathematical Physics

Abstract: I will try to recall some ideas from Linear algebra, so to speak in a simple way, and hopefully understandable, to the principles of Quantum Mechanics, as understood by modern mathematical physics, avoiding complicated questions and technical computations.

- talk 2
**Parkinson’s disease (PD): Generalities and example of a possible genetic cause of PD: Miro1 (Phd Thesis)**(Axel Chemla)

Domain: Biology,Neuroscience

Abstract: First of all, I will introduce Parkinson’s Disease -the most common neurodegenerative movement disorder- coming from the symptoms seen in the patient to the more cellular and genetic effects and causes. Then, the model that we use in my lab will be explained: Neurons coming from patients’ skin biopsies, and induced pluripotent stem cells, and what biological aspects we focus on in the group I am working in. Following this, a concrete example will be given with the gene I am studying during these doctoral studies. Lastly, the final goal of our lab, coming from bed to benchside and back to the hospital bed, will serve as a conclusion for this talk.

#### 4.1.2 (12:30pm-13:30pm, room MNO 1.050)

chair: Jeroen VAN WIER

food: Alessia SCIORTINO

- talk 1
**Artin's conjecture on primitive roots**(Pietro SGOBBA)

Domain: Number theory

Abstract: Let p be a prime number and let Z/pZ be the set of possible remainders obtained by division by p, that is {0,1,2,…,p-1}. Elements of this set can be added and multiplied. One can show that the multiplicative set (Z/pZ)*={1,2,3,…,p-1} has a special element g, called primitive root, which is such that every element b of (Z/pZ)* can be written as a power of g, say b=g

^{k}for some integer k. In other words, the set (Z/pZ)* is given by {1,g,g^{2,g}^{3,…,g}^{(p-1)}}. A question arises naturally. Is there an integer n in Z such that its reduction n mod p is a primitive root for infinitely many primes p? We notice that if such a number exists, it cannot be 0,1,-1 clearly, and not a square either. In fact, Artin conjectured in 1927 that the answer to this question is yes for every other integer. However this has not been proven (or disproven) yet, except for a proof of Hooley from 1967 which relies on a generalised version of the Riemann hypothesis. - talk 2
**Characterising Critical Transitions in Melanoma Cells**(Tijana RANDIC)

Domain: Systems and Molecular Biomedicine

Abstract: Utilizing sc-RNA sequencing and corresponding bioinformatical tools to unravel underling mechanisms causing switch from healthy skin to early melanoma, as well as BRAFi sensitive melanoma cells transitions towards BRAFi resistant melanoma cells.

#### 4.1.3 (12:30pm-13:30pm, room MNO 1.050)

chair: Tieu Long MAI

food: Pietro SGOBBA

- talk 1
**The Local-Global Principle**(Sebestiano Tronto)

Domain: Number Theory

Abstract: The most elementary objects in Mathematics are the natural numbers (0, 1, 2…). The most elementary operations between natural numbers are addition and multiplication. With these operations, we can form equations and ask ourselves: which natural numbers are a solution for this equation? Despite being an elementary and very old question, there is no general method to answer this question, and depending on the equation the problem can be very hard: it took Mathematicians hundreds of years (and thousands of pages) to prove that equations of the form X

^{n}+ Y^{n}= Z^{n}have no solutions for n bigger than 2 (except the "trivial" ones). In this talk we will introduce the so-called "Local-Global Principle": the idea is that looking at our equation modulo different prime numbers ("locally") might tell us something about the "global" solutions. This is a modern topic of research in Number Theory and Algebraic Geometry. - talk 2
**Computational System Biology – Network Inference**(Celine BARLIER)

Domain: Computational Biology

Abstract: In this talk, I will explain what is computational system biology by focusing on basics of gene regulatory networks and the importance of the system-level. After explaining the aims of my PhD topic, I will present in an easy way one of my current group project aiming to infer transcriptional regulatory core from single-cell data : Single-cell gene expression profiling can resolve cell heterogeneity and thus uncover specific cell subpopulations. This high-resolution data is versatile for building cell subpopulation specific gene regulatory networks (GRNs). However, despite the existence of computational methods to infer GRNs, they suffer from numerous limitations and often perform poorly. Here we propose a new computational method to infer transcriptional regulatory core of cell types and subpopulations from single-cell data. Our algorithm is based on a pair-wise distance measure in the gene expression landscape and uses Chip-seq data as a backbone to infer cell subpopulation specific transcriptional regulatory core. We evaluated the performance of our method by using different datasets and by comparing against performances of existing GRN inference methods. Based on a rigorous assessment of the performance of all these methods, we show that our method which is based on distance metric outperforms the existing ones for the inference of cell subpopulation specific transcriptional regulatory core.

#### 4.1.4 (12:30pm-13:30pm, room MNO 1.050)

chair: Axel Chemla

food: Massimo Notarnicola

- talk 1
**DISTRIBUTED COMPUTING: PAST, PRESENT AND THE FUTURE**([ABDALLAH ALI])

Domain: COMMUNICATIONS AND DISTRIBUTED COMPUTING

Abstract: WITH THE MASSIVE DEVELOPMENT AND THE EVOLUTION IN DISTRIBUTED COMPUTING TECHNOLOGY, IT HAS BECOME POSSIBLE FOR MOST USERS TO ENJOY THE BENEFITS AND THE SERVICES OF HIGH PERFORMANCE OR MULTIPURPOSE COMPUTING. TO ACHIEVE THE PREVIOUS FACILITIES TO THE USERS, IT CAN BE ACHIEVED THROUGH CLOUD COMPUTING. CLOUD COMPUTING IS A TERM WHICH REFERS TO PROVIDING AND DELIVERING HOSTED SERVICES OVER THE INTERNET. HOSTED SERVICES ARE LIKE PLATFORM AS A SERVICE (PAAS), SOFTWARE AS A SERVICE (SAAS), ETC. CLOUD COMPUTING IS CONSIDERED AS ONE OF THE EMERGING AREAS OF COMPUTER SCIENCE IN RECENT TIMES. IT IS PROVIDING EXCELLENT FACILITIES TO BUSINESS ENTREPRENEURS BY FLEXIBLE INFRASTRUCTURE. ALTHOUGH CLOUD COMPUTING IS FACILITATING THE INFORMATION TECHNOLOGY INDUSTRY, THE RESEARCH AND DEVELOPMENT IN THIS ARENA ARE YET TO BE SATISFACTORY. IN THIS SESSION, AN ADVANCED INTRODUCTION TO CLOUD COMPUTING WILL BE PRESENTED. THIS INTRODUCTION COVERS THE MOST IMPORTANT TOPICS IN CLOUD COMPUTING. THE INTRODUCTION WILL BE STARTED BY THE HISTORY OF COMPUTING AND ENDS WITH THE FUTURE OF CLOUD COMPUTING. IT COVERS ALL THE CLOUD COMPUTING TECHNOLOGIES, BENEFITS, CHALLENGES AND SERVICES.

- talk 2
**DIFFERENTIAL PRIVACY**([BALAZS PEJO])

Domain: COMPUTER SCIENCE

Abstract: SHORTLY AFTER ITS INTRODUCTION IN 2006, DIFFERENTIAL PRIVACY BECAME THE FLAGSHIP OF DATA PRIVACY DEFINITIONS. SINCE THEN, A LARGE NUMBER (50+) OF VARIANTS AND EXTENSIONS WERE PROPOSED TO ADAPT IT TO DIFFERENT SCENARIOS AND ATTACK MODELS. IN THIS TALK, AFTER A SHORT REVIEW OF MY PHD TOPIC, I PROPOSE A SYSTEMATIC TAXONOMY OF THESE RELAXATIONS INTO SEVEN DIMENSIONS, DEPENDING ON WHICH ASPECT OF THE ORIGINAL DEFINITION IS MODIFIED.

#### 4.1.5 (12:30pm-13:30pm, room MNO 1.040)

chair: Alexey Kalugin

food: Sebastiano Tronto

- talk 1
**Nodal portraits of chaotic billiards**(Massimo Notarnicola)

Domain: Mathematics.

Abstract: At the end of the 18th century, Ernst Chladni, a physicist and musician, made an interesting discovery: he observed that when he excited a metal plate with the bow of his violin, he could hear sounds of different frequency. The plate was fixed only at its center, and when Chladni put some sand on it, then for each frequency a curious pattern appeared, today known as Chladni figures. Some time later, Kirchhoff pointed out that these patterns correspond to nodal sets of eigenfunctions of the biharmonic operator. Since then, the study of nodal sets of Laplacian eigenfunctions has gained substantial attraction and found numerous applications, e.g. in quantum mechanics, statistical physics, etc. In this talk, we introduce planar random waves and discuss why this model, despite being simple, is of importance in a currently developing stream of research in probability theory.

- talk 2
**Don’t say things that you might regret: formal verification of communication protocols**(Itzel Vazquez Sandoval)

Domain: Logic, formal methods, cryptography

Abstract: Whatsapp? Facebook messenger? Telegram? No matter which app you use to communicate, would you care about a stranger reading your conversations? What about someone telling everybody who you have been talking with in the last month, for how long and where your favorite contacts are located? Or even if you hide it, what about a gossiper accessing all the history of all your chats? Journalists that share revealing secrets about governments or about the organizations that control the world definitely do care. In this talk we will introduce a technique based on formal methods that intends to make us aware about the guarantees and lacks of digital communications nowadays; we adopt the mindset of someone eager for information, in order to explore, find and hopefully prevent attacks to privacy.

#### 4.1.6 (12:30pm-13:30pm, room MNO 1.040)

chair: Abdallah Ibrahim

food: LI CHEN

- talk 1
**Studying the prodromal phase of Parkinson’s Disease**(Alessia SCIORTINO)

Domain: Bioinformatics

Abstract: Parkinson’s Disease is the second most common neurodegenerative disease after Alzheimer. The increasing age of the population worldwide and the subsequent increasing number of patients per year, it becomes more and more crucial to understand the basics of the pathology. After giving some basic definitions on what PD is and of its clinical and molecular characteristics, I will explain why it is important to study the prodromal phase (namely the phase that precedes PD clinical motor symptoms onset) and I will present the outline of my project. Some preliminary results will also be included.

- talk 2
**Time-Sensitive Networking (TSN): A quiet revolution in the industry**(Tieu Long MAI)

Domain: Information Sciences

Abstract: We know that a break in a modern car is not a real physical break, it is actually an electronic device that sends the "break" data to the controller. The data is transferred in the car by the Ethernet network. Normally, Internet users find that sometimes the network is slow, for instance, watching Youtube videos with delay. It is however not acceptable with the "break" data in the car! Time-Sensitive Networking (TSN) provides a solution that modifies the Ethernet network to guarantee that the "critical" data is transferred without delay. TSN is expected to be the new standard in many industries, in particular cars or satellites.

#### 4.1.7 (12:30pm-13:30pm, room MNO 1.040)

chair: Antonio Di Maio

food: Eduardo IBARGUENGOYTIA

- talk 1
**Quantum Deniability**(Jeroen van Wier)

Domain: Cryptography, Theoretical Computer Science

Abstract: Consider the following setting: Alice sends an encrypted invitation for a party to Bob. Eve, the "friend" of Alice that will not be invited, intercepts this message but cannot decipher it. Being the annoying friend she is, Eve goes to Alice and demands to know what the message says. This is an example where deniability can be very useful. For many forms of encryption, Alice would have to tell Eve exactly what the message says in order to convince her because Eve can check this. But if the encryption is deniable, then Alice can give Eve a completely fake message and still convince Eve. In this talk, we will go over the basics of encryption, and then discuss different forms of deniability. If time permits, we will also discuss what new challenges the "quantum" version of deniability poses.

- talk 2
**An exposition on Hyperbolic Geometry and beyond**(Diptaishik Choudhury)

Domain: Mathematics

Abstract: In this talk, we will explore hyperbolic geometry in an intuitive way. If time permits, we can then proceed to understand some peculiarities of spaces adhering to this kind of geometry.

### 4.2 seminar 2018:

date | room | speaker 1 | speaker 2 | chair | food |
---|---|---|---|---|---|

1.020 | German Preciat Gonzalez | Yujuan Gui | Alexey Kalugin | Assar Andersson | |

1.040 | Susana Martinez | Maria Fernandes | Loulou Peisl | Emiliano Torti | |

1.040 | Loulou Peisl | Emiliano Torti | Yujuan Gui | Luca Notarnicola | |

1.040 | Antonio di Maio | Guenda Palmirotta | Ines Pinto Gouveia | Ines Pinto Gouveia | |

1.020 | Alexey Kalugin | Assar Andersson | Antonio di Maio | Susana Martinez | |

1.040 | Ines Pinto Gouveia | Luca Notarnicola | Guenda Palmirotta | German Preciat Gonzalez |

#### 4.2.1 (1pm-2pm, room 1.040)

chair: Guenda Palmirotta

food: German Preciat Gonzalez

- talk 1
**Byzantine fault tolerance using shared memory objects**(Ines Pinto Gouveia)

Domain: Distributed Byzantine fault tolerance

Abstract: Byzantine fault tolerance (BFT) algorithms for message-passing systems have been an active research line throughout the years. However, few are the studies that consider BFT and consensus approaches through shared memory. Seeking to improve on limitations from previous works, recent attempts targeted the use of policy-enforced augmented tuple spaces. The aim was to reduce the amount of objects and operations required, making shared memory BFT practical for real systems.

- talk 2
**Key Exchange Protocols in Cryptography**(Luca Notarnicola)

Domain: Mathematics - Cryptography

Abstract: In 1976 Diffie and Hellman constructed a protocol that allows two people to securely agree on a common shared key by exchanging data over an insecure channel. The first goal of this talk is to introduce their construction and then to move to the following question, which interestingly remains open: Can this protocol be generalized to more than 2 users? This leads to the study of the so-called cryptographic multilinear maps, which we will then briefly introduce.

#### 4.2.2 (12:15pm-1:15pm, room 1.020)

chair: Antonio di Maio

food: Susana Martinez

- talk 1
**Periods and multi zeta values**(Alexey Kalugin)

Abstract:

Values of the zeta function for even numbers are well known, due to the work of Euler (1735). All of them are some multiplicities of pi and thus are transcendental numbers. From the other hand values of zeta functions for odd numbers are completely mysterious. Famous Transcendence conjecture states that all of them are algebraically independent, but this conjecture was out of the reach for centuries. I am going to explain recent breakthrough towards the proof of this conjecture.

- talk 2
**Value of risk and the Kelly criterion**(Assar Andersson)

Abstract:

In this talk we are going to discuss a simple mathematical model to put a value of a risk. This can be used to help you determine how much money you should bet on your favourite football team, or whether you should buy an insurance or not.

#### 4.2.3 (12:15pm-1:15pm, room 1.040)

chair: Ines Pinto Gouveia

food: Luca Notarnicola

- talk 1
**Software-Defined Vehicular Networking**(Antonio di Maio)

Abstract: Every year, billions of kilometers are driven by cars on the roads of the world. Drivers spend thousands of hours per year in their vehicles, stuck (and angry) in traffic. Danger and inefficiencies on the road urge researchers to investigate how to guarantee safer and more enjoyable journeys to drivers and passengers. In this talk, we will see how wireless and cellular data networking can be embedded in cars and how it will play a role in saving lives, time and stress of millions of drivers.

- talk 2
**PDE's in Action – Modelling the brain with “Lego bricks”**(Guenda Palmirotta)

Domain: Mathematics

Abstract: Surgeons often operate in the dark. They have a limited view of the surface of the organ, and can typically not see what lies hidden inside. Quality images can routinely be taken prior to the surgery, but as soon as the operation begins, the position of the surgeon’s target and risky areas he must avoid, continuously change. This forces practitioners to rely on their experience when navigating surgical instruments to, for example, remove a tumor without damaging healthy tissue or cutting through important blood supplies. To solve this problem, Dr. Stéphane Bordas and his team developed a computational method that could be used to guide surgeons during brain surgery.

The aim of this talk is to explain how PDE's (Partial Differential Equations), more precisely FEM (Finite Element Methods), could be applied to solve such kind of problem(s).

#### 4.2.4 (12:45pm-1:45pm, room 1.040)

chair: Yujuan Gui

food: Ines Pinto Gouveia

- talk 1
**Elucidation of previously uncharacterized metabolites involved in human-microbial cross-talk**(Loulou Peisl)

#UnMetXTalk

Domain: Systems Biology/Human Microbiome

Abstract: UnMetXTalk, a joint project between the Eco-Systems Biology and Environmental Cheminformatics groups, aims to investigate the interplay between dietary fibre, gut microbiota, produced metabolites, and their effect on human cell physiology. To achieve this, a microfluidics-based human-microbial co-culture system that mimics the gastrointestinal human-microbe interface (HuMiX) will be combined with mass spectrometry (MS) methods and state-of-the-art computational approaches. In particular, UnMetXTalk aims to systematically resolve small molecule exchange between human epithelial gut cells and a specific microbiota and to identify unknown metabolites.

- talk 2
**A mathematical model for the dynamics of a prey-predator system**(Emiliano Torti)

Abstract: In this talk i will speak about one of the most famous example of mathematical model used in biology. It is the one created by Lotka and Volterra around 1925 to describe the populations change through time in a predator-prey system.

#### 4.2.5 (12:15pm-1:15pm, room 1.040)

chair: Loulou Peisl

food: Emiliano Torti

- talk 1
**Critical transitions in microbial communities**(Susana Martinez)

Domain: Systems Biology

Abstract: Microbial communities are made up of multiple populations of microorganisms that share a common habitat, at a particular time. They are ubiquitous, complex and dynamic systems that play important roles in natural environments, human health and biotechnological processes. This project involves the identification and understanding of critical transitions within microbial communities, in particular, this project focus on the dynamics of bacteriophage (and other invasive genetic elements) and bacterial host interactions through the use of time-resolved, multi-omics datasets derived from different environments, i.e biological wastewater treatment plant and gut microbiome.

- talk 2
**Privacy-preserving alignment**(Maria Fernandes)

Domain: Genomic privacy/Bioinformatics

Abstract: Next-generation sequencing technologies advances allowed the great availability of sequencing data. Sequencing thousands of human genomes has enabled breakthroughs in many areas, among them precision medicine, the study of rare diseases, and forensics. However, mass collection entails enormous privacy risks if not protected to the highest standards.

#### 4.2.6 (12-1pm, room 1.020)

chair: Alexey Kalugin

food: Assar Andersson

- talk 1:
**Novel constraint-based modelling techniques for experimental design**(German Andres Preciat Gonzalez)

abstract: Biological systems such as organisms and cells are highly organised in their structure and function. They evolve over time by adapting to their environment, and therefore, their behaviour is difficult to predict from the properties of individual parts. Systems biology studies the components of biological systems using an integrated approach to analyse at complex data from multiple experimental sources (genomics, transcriptomics, proteomics, metabolomics, etc.) to understand how the interactions of components give rise to their biological behaviour and function. In this talk, will be discussed novel approaches to study a biological system using its mathematical representation.

- talk 2:
**Genome-wide association studies (GWAS) identifying risk Loci of Parkinson’s disease**(Yujuan Gui)

domain: biology

abstract: In recent years, there are many biobanks being established to collect human genomic data, which facilitates genome wide association studies (GWAS) to identify correlated or causal genes/loci related to complex trait diseases like the Parkinson’s disease (PD). The identified risk loci are potential targets for clinical research. The basic idea of GWAS study is to take use of many people’s genomic data, find the SNPs and associate these SNPs to the trait of interest.

### 4.3 seminar 2017:

date | speaker 1 | speaker 2 | chair | food |
---|---|---|---|---|

Andrew Yarmola | Tamara Bintener | Qingyuan Hu | Ivan Nourdin | |

Guangqu Zheng | Qingyuan Hu | Tamara Bintener | Diu Tran | |

Hao Chen | - | Hussein Rappel | Anna Vidotto | |

Anna Vidotto | Hussein Rappel | Diu Tran | Guangqu Zheng | |

Diu Tran | Jimmy Devillet | Guangqu Zheng | Simon Campese |

chair: Guangqu Zheng

food: Simon Campese

- talk 1:
**Self-similar stochastic processes**(Diu Tran)

domain: *Math-Probability and Statistics

abstract: We will introduce the story about self-similar processes and their applications in our real life. No mathematical background is needed.

- talk 2:
**Quasi-arithmetic means**(Jimmy Devillet.)

domain: Mathematics

abstract: Already discovered and studied by the ancient Greeks, the concept of mean has given rise today to a very wide field of investigation with a huge variety of applications. Indeed, for instance, means are ubiquitous in statistics.

In this talk we focus on the class of quasi-arithmetic means on a real interval. Indeed, we introduce these functions and observe how they are related to the arithmetic mean. Moreover, we provide a very useful tool in order to compare such means and apply it in order to reprove the well-known geometric-arithmetic mean inequality. Finally, we provide an axiomatic characterization of quasi-arithmetic means. All the introduced notions and results will be illustrated with examples.

- talk 1:

chair: Diu Tran

food: Guangqu Zheng

- talk 1:
**Normal approximation in a simple example**(Anna Vidotto)

domain: Mathematics, Probability abstract: We will show an application of the Central Limit Theorem to a simple example regarding a dice game.

- talk 2:
**Bayesian inference for material parameter identification**(Hussein Rappel)

domain: Computational science, Parameter identification and Mechanics of materials

abstract: Material parameter identification based on error minimisation (e.g. least squares) is frequently used in the field of mechanics. These conventional error minimisation methods do not explicitly account for uncertainties and noise of the experimental devices. An alternative identification method is the Bayesian approach which leads to a probabilistic estimation of material parameters. It can therefore not only characterise the most probable parameter values, but also the uncertainty of these values. This talk will be about Bayesian inference for parameter identification in elastoplasticity.

- talk 1:

chair: Hussein Rappel

food: Anna Vidotto

- talk 1:
**Periodic minimal surfaces**(Hao Chen)

domain: Mathematics and Material science

abstract: Originally motivated by bubble films, the theory of minimal surface naturally lies in the intersection of mathematics and physics. Periodic minimal surfaces are minimal surfaces with the symmetry of crystals. They have been observed in biological membranes, lyotropic liquid crystals, and many other natural or artificial materials. Mathematically, periodic minimal surfaces involve deep theories of complex analysis, functional analysis and differential geometry. It is then not a surprise that physicists, chemists and mathematicians have been working together to discover, construct and analyze periodic minimal surfaces. I will review this exciting history of collaboration, including the most recent development in which the speaker played a role. The talk will be accompanied with interesting anecdotes and nice 3D pictures.

- talk 1:

chair: Tamara Bintener

food: Diu Tran

- talk 1:
**Computing the approximate expectation and error control**(Guanqu Zheng)

domain: Probability theory and related fields.

abstract: We will briefly introduce Stein’s method of normal approximation.

- talk 2:
**Nitsche method for interface problems in isogeometric analysis**(Qingyuan Hu)

domain: Computational Solid Mechanics

abstract: Isogeometric analysis (IGA) is nowadays taken as a bridge between computer-aided design (CAD) and computer-aided engineering (CAE), instead of finite element method (FEM). However one drawback of IGA is the non-interpolated property of the adopted shape functions, i.e. non-uniform rational basis spline (NURBS). In this talk I will introduce the Nitsche formulation as a powerful tool, to handle some practical details, including, how to implement Dirichlet boundary conditions, how to couple dis-matched meshes, how to simulate contact.

- talk 1:

chair: Qingyuan Hu

food: Ivan Nourdin

- talk 1:
**Computational techniques for hyperbolic 3-manifolds**(Andrew Yarmola)

domain: geometric structures on surfaces, Kleinian groups, and computational topology and geometry

dbstract: In this talk, I will introduce the notion of a hyperbolic 3-manifold and explain some aspects that are particularly well suited to computational approaches. My focus will be on understanding hyperbolic 3-manifolds of low cusp volume and the interplay between algebraic, topological, and computational techniques that lead to a classification of the infinite families of hyperbolic 3-manifolds of cusp volume.

- talk 2:
**Methods in metabolic modelling and their applications**(Tamara Bintener)

domain: Systems Biology

abstract: In the past years, metabolic modelling has gained and increasing interest to study metabolic diseases as well as cancer. With the recent availability of genome-scale metabolic reconstructions and different model building algorithms, it is now possible to create context-specific models. Metabolic modelling approaches such as flux balance analysis can be used to predict the flux of a metabolite or to find essential genes. Furthermore, the comparison of diseased and healthy context-specific models can help elucidate metabolic differences between the models. Here, I will briefly explain the mathematical background of the algorithms and approaches in metabolic modelling.

- talk 1: