MIDASMIDAS
The 8ht Workshop on MIning DAta for financial applicationS
September 22, 2023 - Turin, Italy
http://midas.portici.enea.it
in
conjunction with
ECML-PKDD 2023
European Conference on Machine Learning and Principles and Practice of Knowledge Discovery
September 18-22, 2023 - Turin, Italy
We invite submissions to the 8th MIDAS Workshop on MIning DAta for financial applicationS, to be held in conjunction with ECML-PKDD 2023 - European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases.
Like the famous King Midas, popularly remembered in
Greek mythology for his ability to turn everything he
touched with his hand into gold, we believe that the
wealth of data generated by modern technologies, with
widespread presence of computers, users and media
connected by Internet, is a goldmine for tackling a
variety of problems in the financial domain.
The MIDAS workshop is aimed at discussing challenges, potentialities, and applications of leveraging data-mining tasks to tackle problems in the financial domain. The workshop provides a premier forum for sharing findings, knowledge, insights, experience and lessons learned from mining data generated in various application domains. The intrinsic interdisciplinary nature of the workshop constitutes an invaluable opportunity to promote interaction between computer scientists, physicists, mathematicians, economists and financial analysts, thus paving the way for an exciting and stimulating environment involving researchers and practitioners from different areas.
PAST EDITIONS
We encourage submission of papers on the area of data mining for financial applications. Topics of interest include, but are not limited to:
WORKSHOP FORMAT
The ECML-PKDD 2023 conference - and all its satellite events, including the MIDAS workshop - will adopt a "hybrid" format, with both in-person and remote attendance allowed. However, in order to maximize engagement and physical presence at the conference, remote attendance - with the associated remote registration fee - is considered an option for non-presenting attendees only.
INVITED SPEAKERS
Dr. Roberto Pellungrini, Scuola Normale Superiore
Title.
Explainable AI and the financial sector.
Abstract.
Explainable AI is a relatively novel frontier of AI that enables the development of ML system that are transparent and trustworthy. These properties are fundamental for a sector like the financial one, where algorithmic decision have a strong impact on peoples wellbeing.
Explainability can enable a more fruitful communication between experts in the field and customers, to understand the decision of the algorithms and understand what actions to put into place to change that decision.
In this talk I’m going to give an overview of Explainable AI techniques for tabular data, most common in this sector, and I’m going to illustrate some of the Explainable AI literature focused on problems related to the financial sector, such as Churning, Credit Scoring and Portfolio Management.
Bio.
Roberto Pellungrini is a Research Fellow ad Scuola Normale Superiore, Classe di Scienze.
He earned his PhD in Computer Science at the Department of Computer Science, University of Pisa in 2020, with a thesis on Data Privacy.
His current research interest are Explainable AI and Learning Paradigms for Hybrid Decision Making Systems.
Prof. Edoardo Serra, Boise State University
Title.
Advancements in Graph Representation Learning: Integrating Structural Aspects into Temporal Graph Representations for Financial Transactional Data Applications.
Abstract.
Over recent years, there has been a growing interest in the field of graph representation learning methods. These techniques are designed to generate numerical representations of various components within a graph, such as nodes, edges, and the entire graph itself. Graph representation learning methods can be broadly categorized into two primary groups: those that aim to retain the connectivity information of nodes and those that aim to preserve the structural information of nodes.
Connectivity-based methods are focused on capturing the relationships between nodes, resulting in nodes that are connected being closer together in the latent space they generate. On the other hand, structural-based methods concentrate on preserving the structural characteristics of nodes, bringing nodes with similar structures closer together in the resulting space.
Extending structural graph representation learning to temporal graphs poses a challenging problem, but it holds significant importance, especially in the context of financial transactional data. This presentation will provide an overview of structural and proximity-based representation learning, emphasizing the significance of structural-based approaches. We will also explore extensions of graph representation learning to temporal graphs and discuss their application to financial transactional data.
Bio.
Dr. Edoardo Serra holds the position of Associate Professor within the Department of Computer Science at Boise State University, concurrently serving as a Senior Researcher with a Joint Appointment at the Pacific Northwest National Laboratory (PNNL). He earned his Ph.D. in Computer Science Engineering from the University of Calabria, Italy, in 2012. Dr. Serra's research expertise lies in the domains of Graph Representation Learning, Robust and Machine Learning, and the practical application of Machine Learning and Artificial Intelligence for National Security.
ACCEPTED PAPERS
Elena Tiukhova, Adriano Salcuni, Can Oguz, Marcella Niglio, Giuseppe Storti, Fabio Forte, Bart Baesens, Monique Snoeck - "Boosting Credit Risk Data Quality using Machine Learning and eXplainable AI Techniques" [PRESENTATION]
PROGRAM