<div dir="ltr"><pre style="color:rgb(0,0,0)">***************************************************************
CALL FOR PAPERS (ESANN 2025)
European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning
<a href="https://www.esann.org/" target="_blank">https://www.esann.org/</a>
More details on the Special Session "Foundation and Generative Models for Graphs":
<a href="https://www.esann.org/ESANN2025specialsessions#graphs" target="_blank">https://www.esann.org/ESANN2025specialsessions#graphs</a>
Bruges, Belgium, 23-25 April 2025
***************************************************************
****** Call For Papers ******
Graphs are versatile abstractions that model complex systems of interacting entities, where interactions imply functional and/or structural dependencies among them. Molecular compounds and social networks are two particularly relevant instances of graph-structured data: the former can be viewed as a system of interacting atoms, where chemical bonds are influenced by factors like inter-atomic distances and atomic energies/forces depend on long-range electrostatic interactions. Instead, social networks manifest in terms of user-user and user-content interactions, where content is multimodal and includes pictures, movies, and songs. Other practical examples include the encoding of symmetries and constraints in combinatorial optimization problems, which serve as proxies for apriori knowledge.The field of deep learning for graph-structured data focuses on adapting deep learning techniques, such as convolutional operators, to the analysis and processing of graphs. Recent advancements in foundation and generative models for non-graph data have significantly impacted businesses, with an unprecedented speed of adoption in production environments. It is therefore natural to ask ourselves whether such advances can be readily transferred to the domain of graphs, and what the best way to do so would be, opening new avenues for research and applications. In this respect, there are high expectations in areas such as drug discovery, where the ability to generate molecular constraint-preserving graphs with desired properties might reduce the humongous amount of money and compute time needed to screen candidate drugs. This special session is an excellent opportunity for the machine learning community to gather together and host novel ideas, showcase potential applications, and discuss the new directions of this remarkably successful research field.<br></pre><pre style="color:rgb(0,0,0)">****** Important Dates ******
Paper Submissions: 20 November 2024
Paper Acceptance Notifications: 24 January 2025
Conference: 23-25 April 2025
****** Topics ******
This special session aims to gather valuable contributions and new findings in the field of deep learning for graphs. It will also emphasize the integration of foundation and generative models to enhance the creation and manipulation of graph-structured data, unlocking new possibilities in areas such as molecular and material design. In particular, we look for contributions in the following areas:<br>-> Architectures for foundation models operating on graphs;
-> Graph generation (e.g., probabilistic models, variational autoencoders, normalizing flow, diffusion models);
adversarial learning, etc.)ù
-> Graph representation learning;
-> Graph structure learning and relational inference;
-> Graph coarsening and pooling in graph neural networks;
-> Theory of graph neural networks (e.g., expressive power, learnability, negative results);
-> Learning on complex graphs (e.g., dynamic graphs, heterogeneous graphs, and spatio-temporal data);
-> Anomaly and change detection in graph data;
-> Randomized neural networks for graphs (e.g., reservoir computing);
-> Recurrent, recursive, and contextual models;
-> Scalability, data efficiency, and training techniques of graph neural networks;
-> Tensor methods for structured data;
-> Graph datasets and benchmarks.
****** Session Organisers ******
-> Davide Rigoni (University of Padua)
-> Luca Pasa (University of Padua)
-> Federico Errica (NEC Laboratories Europe)
-> Daniele Zambon (Swiss AI Lab IDSIA, Università della Svizzera italiana)
-> Davide Bacciu (University of Pisa)
-> Stefano Moro (University of Padua)</pre></div>