We welcome contributions on Temporal Networks (also referred to as Dynamic Graphs, Stream Graphs, Time-Varying Networks, Evolving Networks, or Link Streams) applied to any real-world contexts in conjunction with various fields, such as:
Dynamic community detection
Event detection
Signed Networks
Higher-order interactional data
Network Data Collection
Software for temporal graph analysis
Graph machine learning
Link prediction
Modeling fairness and ethics
Network Measures And Metrics
Studies that combine complexity and temporal networks that do not fall within the above are also welcome!
Oral Presentations: Accepted talks will be allocated a 15-minute slot, consisting of 12 minutes for presentation and 3 minutes for audience questions.
Poster Presentations: Accepted posters will be featured in dedicated poster sessions. Additionally, poster presenters will have a 2-minute slot during the pitch talk session to highlight their work. All posters should be self-printed in A0 format (84.1 cm × 118.9 cm, portrait orientation). Presenters of the pitch talks will be asked to send the slides (in pdf format) in advance, to allow a perfect flow of the session.
(Deadline for sending the pitch talks : 1 June 2025)
We will also present the Best Talk/Poster Award among all satellite contributions.
Selected contributions will be invited to submit to a dedicated Special Issue, Evolution of Networks, for the Applied Network Science journal.
Program
Posters
(Listed without any specific ranking or preference/Same order will apply in the pitch session)
Annalisa Caligiuri: Characterizing the dynamics of unlabelled temporal network trajectories
Tianrui Mao: Estimating nodal spreading influence using partially observed temporal network
Kevin Teo: Periodicity detection in the global shipping network using dynamic mode decomposition
Raphaël Romero: Multi-Scale Detection in Temporal Networks via Low-Rank Point Process Projections
Alessio Catanzaro: Fast Networks and Slow Observers: Lessons in Reconstruction from Renormalization
Pau Esteve: Flight2vec: Event embedding for flight delay prediction
Prathyush Sambaturu: Modeling and Analyzing H5N1 Spread in U.S. Cattle Using Temporal Networks
Tomomi Kito: Understanding idea formation in teams: a temporal network approach
Tassilo Schwarz: Mind the Memory: Emergence of higher-order dynamics and consistent statistical entropy production rate estimation in continuous-time complex systems
Keynote Speakers
Naoki Masuda
State University of New York at Buffalo, USA
Node-state dynamics view of temporal networks
A salient feature of empirical temporal network data is heavy-tailed distributions of inter-contact times. This phenomenon has caught a lot of attention since early days of temporal network research, and not only impacts dynamical processes occurring on networks but also gives us insight into behavior of nodes generating such statistics. I introduce modeling of this phenomenon in which each node is assigned a latent variable that is dynamic in different ways. This approach is interpretable, is analytically tractable using random walk theory (including the case where one looks at autocorrelation function as opposed to distributions of inter-event times), and seeds various research questions (e.g., dynamics on the generated temporal networks, inference, relating to other node-variable models of networks), as I discuss in the present talk.
McGill University, Canada
Deep Learning Practices in Temporal Networks
Many real-world networks contain crucial time-domain information, which temporal graphs capture through temporal events, making them valuable across various domains where time-varying patterns are essential. Building on the success of graph representation learning in static graphs, researchers have developed dynamic graph representation models to leverage temporal information in dynamic graphs. These models have shown superior accuracy in downstream tasks such as temporal link prediction and dynamic node classification, outperforming static approaches and traditional methods on diverse networks like social networks, traffic graphs, and knowledge graphs. However, dynamic graph models face significant challenges in data efficiency, resource efficiency, and evaluation.
This talk introduces the Temporal Graph Benchmark (TGB), a collection of challenging and diverse datasets designed for realistic, reproducible, and robust evaluation of machine learning models on temporal graphs. TGB includes tasks for dynamic link and node property predictions, as well as an automated pipeline from data loading to model evaluation. It also features novel datasets for temporal knowledge graphs and temporal heterogeneous graphs. The talk aims to benefit researchers and practitioners by discussing state-of-the-art dynamic graph representation learning models, focusing on algorithms, frameworks, and tools. Additionally, it will explore promising yet under-explored research directions in efficient temporal graph learning.
Central european university, Austria
Static representations of temporal networks and what they are good for
Temporal networks are commonly used to represent systems where connections between elements are active only for restricted periods of time, such as telecommunication, biochemical reactions or social networks. The time-varying nature of such interactions determine several network properties, like valid temporal paths, that in turn influence the emergence of any macroscopic phenomena on the time-varying structure. Nevertheless, the actual representations of temporal networks hardly allow the effective computation of these dynamical network properties. In this talk we will take an overview about static representations of temporal networks, especially focusing on event graphs, that allow to describe a sequence of time-varying interactions as static structures that can be further analysed effectively. The proposed representations open an avenue to the lossless description and computationally efficient characterisation of very large temporal networks and ongoing dynamics processes relevant to understand spreading phenomena.
Aalto University, Finland
Perspective on Temporal Network Science
The recursive idea that things get their functionality from how they are connected to other things connects many branches of the social, natural, and formal sciences, From the structuralism of last-century social science via network science to today's graph-based machine learning. If one also knows when the interactions happen, one should have more information and get a better understanding of the system—but how should one extract, visualize, and theorize this type of data? I will present how these ideas have developed throughout the last 150 years and connect them to my own research on temporal networks of social interactions, human mobility, network epidemiology, link predictions, and more.
.
Location and general information
The satellite will take place at the same venue as the Netsci 2025 conference, Maastricht University, Netherlands.
Please remember that in order to participate in TENET2025, you need to register for NetSci2025.
For other information, such as VISA requirements, please check the webpage of the main conference.
Yasaman Asgari
University of Zurich, Switzerland
Alessia Galdeman
IT University of Copenhagen, Denmark
Andrea Failla
University of Pisa, Italy
Salvatore Citraro
CNR-ISTI, Italy
Remy Cazabet
Univ. Lyon 1, France
Alexandre Bovet
University of Zurich, Switzerland
Nicola Pedreschi
University of Bari, Italy
Cheick Ba
Queen Mary University of London, UK
Sabrina Gaito
Università degli Studi di Milano, Italy
Renaud Lambiotte
Oxford University, UK
Alessia Antelmi
University of Turin
Antonio Longa
University of Trento
Arianna Pera
IT University of Copenhagen
Dorian C. Quelle
UZH
Francesco Cauteruccio
University of Salerno
Giulia Cencetti
CNRS
Giulio Rossetti
CNR-ISTI
Lucio La Cava
University of Calabria
Manuel Dileo
University of Milan
Shazia Ayn Babul
University of Oxford
Timothy LaRock
University of Oxford
For any inquiries, feel free to reach out to us at tenet.netsci@gmail.com.