The School of the NetSci 2022 conference is inspired by two topics, higher-order interactions, and statistical inference.
Higher-order interactions is an emerging field calling for the generalization of network science principles beyond pairwise interactions. By attending the School, you will have a chance to learn about the most recent developments in the higher-order interactions branch of network science, including generalized diffusive processes, synchronization, higher-order motifs, and communities.
Statistical inference on/of networks has been at the heart of all network science branches. At NetSci2022, we highlight recent developments in statistical inference from several angles, from theory to applications. Following school lectures will enrich your arsenal of statistical inference tools with spectral, information-theoretic, and machine-learning methods.

The time shown in the list is the Greenwich Mean Time (GMT+0:00). Please convert it to the time zone where you are going to attend the event.
Some common locations and corresponding current time:

Schedule Overview

Wednesday, July 20
12:00 13:45 Renaud Lambiotte
"Consensus and random walks on higher-order networks"
13:45 15:30  Yuanzhao Zhang
"Synchronization in networks with higher-order interactions"
Thursday, July 21
12:00 13:45 Jan Nagler
"Fundamentals of Networked Dynamical Interactions: A journey from pairwise to higher-order"
13:45 15:30  Federico Battiston
"The organization of higher-order interactions in real-world systems"
Friday, July 22
12:00 13:45 Jie Sun
"Data Science for Network Science:
A Dynamical and Complex Systems Perspective"
13:45 15:30  Yulia R. Gel
"Topological Data Analysis on Complex Networks: Blockchain Transaction Graphs and Beyond"
15:30 17:15 Cristopher Moore
"Statistical inference and community detection in networks"
Saturday, July 23
12:00 13:45 Yanqing Hu
"Applications of Coding on Complex Networks"
Cancelled Pan Li
"Graph neural networks --- its power, its limit and recent applications"

Detailed Program

Renaud Lambiotte

Professor, University of Oxford

Consensus and random walks on higher-order networks

July 20, 12.00 to 13.45 GMT

Abstract: In the last years,  there have been a number of works on the limits of graph-based models to represent interacting systems. Different directions have been proposed in order to enrich the network formalism, leading to the emerging field of higher-order networks. In this lecture, I will focus on random walk and diffusive processes, and discuss their behaviour for different types of higher-order networks,  including temporal networks, non-Markovian networks and hypergraphs. I will also discuss the possibility to use the resulting diffusive processes in order to extract information from the data, for instance via community detection.

Yuanzhao Zhang

Schmidt Science Fellow, Santa Fe Institute

Synchronization in networks with higher-order interactions

July 20, 13.45 to 15.30 GMT

Abstract: Higher-order interactions, through which three or more entities interact simultaneously, are important in many complex systems and have generated great interest in the network science community. In this lecture, I will sample a few rapidly developing directions at the interface between nonlinear dynamics and higher-order networks. A key theme of the lecture will be the effects of nonpairwise interactions on synchronization. For example, when do nonpairwise interactions promote synchronization? What about more complex synchronization patterns such as chimera states? And how do we analyze these synchronization patterns in the presence of nonpairwise interactions?

Jan Nagler

Associate Professor, Centre for Human and Machine Intelligence, Frankfurt School of Finance & Management

Fundamentals of Networked Dynamical Interactions: A journey from pairwise to higher-order

July 21, 12.00 to 13.45 GMT

Abstract: Do higher-order interactions really matter? If yes, when, how and why? I plan to discuss networked examples in signal propagation, percolation, machine learning and causal modeling, with excursions to cosmology, non-self-averaging systems and strong correlates. I will decorate my lecture with related own work on networked dynamical systems, published and unpublished. I begin simple using an electronic black board and ask plain, yet fundamental questions. I hope for an interactive lecture.

Federico Battiston

Assistant Professor, Central European University

The organization of higher-order interactions in real-world systems

July 21, 13.45 to 15.30 GMT

Abstract: The complexity of many biological, social and technological systems stems from the richness of the interactions among their units. Over the past decades, a variety of complex systems has been successfully described as networks whose interacting pairs of nodes are connected by links. Yet, from human communications to chemical reactions and ecological systems, interactions can often occur in groups of three or more nodes and cannot be described simply in terms of dyads. Until recently little attention has been devoted to the higher-order architecture of real complex systems. However, a mounting body of evidence is showing that taking the higher-order structure of these systems into account can enhance our modeling capacities and help us understand and predict their dynamical behavior. In this lecture I will navigate a number of ways to untangle the higher-order organization of real-world systems, from higher-order motifs and filtering procedures to temporality, community structure and signal processing.

Jie Sun

Chief Researcher, Huawei Hong Kong Research Center

Data Science for Network Science: A Dynamical and Complex Systems Perspective

July 22, 12.00 to 13.45 GMT

Abstract: From telecommunication systems to power grids and robotics, complexity is an integral signature in today’s society. Coping with those ever-complex systems call for new ideas and methodology. We start by presenting common challenges in the modeling of complex systems and discussing how data-driven techniques can be useful in mitigating some of the difficulties. Some basic theory from matrices, linear systems, nonlinear dynamics and stochastic processes together with algorithms will be discussed over the talk. We will also survey recent progress on open problems in causal inference, data-driven modeling and optimization of complex networks.

Yulia R. Gel

Professor, University of Texas at Dallas

Topological Data Analysis on Complex Networks: Blockchain Transaction Graphs and Beyond

July 22, 13.45 to 15.30 GMT

Abstract: "Blockchain technology and, in particular, blockchain-based cryptocurrencies offer us information that has never been seen before in the financial world. In contrast to fiat currencies, all transactions of crypto-currencies and crypto- tokens are permanently recorded on distributed ledgers and are publicly available. As a result, this allows us to construct a transaction graph and to assess not only its organization but to glean relationships between transaction graph properties, crypto price dynamics as well as illegal and illicit activities such as emerging ransomware. In this talk we discuss horizons and limitations of what new can be learned from topology and geometry of cryptocurrency transaction graphs whose even global network properties remain scarcely explored. By introducing novel tools based on topological data analysis (TDA), functional data depth, and geometric deep learning, we show that even some subtler patterns in blockchain transaction graphs can provide critical insights for money laundering tracking, price analytics, and market sentiment assessment. If time permits, we also discuss how the TDA concepts can be used to shed new light onto hidden mechanisms behind organization of complex networks, in conjunction with clustering, node classification, and link prediction in social media, cyber-physical systems, and climate-induced insurance analytics"

Cristopher Moore

Resident Professor, Santa Fe Institute

Statistical inference and community detection in networks

July 22, 15.30 to 17.15 GMT

Abstract: "Statistical inference is the art of finding patterns in noisy data. Many modern inference problems are high-dimensional: namely, where both our data and the model we want to fit to it are very large. This includes community detection in networks, topic models in text analysis, and so on. I’ll give an introduction to Bayesian approaches to these problems, where we assume the data was produced by a generative model. I’ll discuss approximate Bayesian methods like belief propagation and expectation maximization. I’ll also discuss connections between these algorithms and spectral approaches. Along the way, I’ll draw analogies with statistical physics, including how free energy barriers can make inference problems hard to solve."

Yanqing Hu

Associate Professor, Southern University of Science and Technology

Applications of Coding on Complex Networks

July 23, 12.00 to 13.45 GMT

Abstract: In this lecture I will introduce the applications of coding method on complex networks. Specifically, in the first hour,  I will establish the relationship between the shortest compressed length of a complex network and its structure predictability. In the second hour, I will show how to identify community structure by encoding the paths of random walks.

Pan Li

Assistant Professor, Purdue University

Graph neural networks --- its power, its limit and recent applications

July 23,cancelled

Abstract: In this talk, we are to introduce graph neural networks for modern learning problems on graphs & networks. We will introduce their basic working mechanism and some important applications. Then, we will introduce a bit more in-depth topic on the power of graph neural networks, particularly on what they can do, what they cannot do, and how to improve their power. If times are allowed, we will introduce some fancy topics regarding graph neural networks, e.g., how to use it to discover patterns in graphs or networks for certain prediction tasks.

School Chair

Maksim Kitsak

Delft University of Technology

Bo Qu

Pengcheng Laborabory

Peng Ji

Fudan University