Embedded Files
II WORKSHOP ON DATA FUSION
FOR ARTIFICIAL INTELLIGENCE
October 19th, 2024
at ECAI 2024 - Santiago de Compostela
The workshop is open to both theoretical and applied works, as long as the focus is on the analysis of the functions used for the information fusion.
The workshop is open to both theoretical and applied works, as long as the focus is on the analysis of the functions used for the information fusion.
Data fusion is crucial in almost every application of artificial intelligence. Classification, image processing, decision-making, big data, or deep learning require collecting and fusing data in appropriate ways to solve specific problems. For this reason, a huge effort is devoted to the development and analysis of data fusion methods.
Data fusion is crucial in almost every application of artificial intelligence. Classification, image processing, decision-making, big data, or deep learning require collecting and fusing data in appropriate ways to solve specific problems. For this reason, a huge effort is devoted to the development and analysis of data fusion methods.
Aggregation functions are one of the most widely used methods in this sense. They are defined as monotone functions with appropriate boundary conditions and include, among others, most of the means or functions such as the product, the minimum, or the maximum. However, in recent years it has been shown that the concept of aggregation function can be too restrictive, as it does not cover some examples that can provide good results in particular applications, as is the case of the mode. Furthermore, some data fusion functions are more general than aggregation functions. For example, the so-called pre-aggregation functions have been proposed to deal with problems ranging from classification to the computational brain, with promising results.
Aggregation functions are one of the most widely used methods in this sense. They are defined as monotone functions with appropriate boundary conditions and include, among others, most of the means or functions such as the product, the minimum, or the maximum. However, in recent years it has been shown that the concept of aggregation function can be too restrictive, as it does not cover some examples that can provide good results in particular applications, as is the case of the mode. Furthermore, some data fusion functions are more general than aggregation functions. For example, the so-called pre-aggregation functions have been proposed to deal with problems ranging from classification to the computational brain, with promising results.
In this workshop, we intend to review the most recent developments in the field of data fusion with uncertainty, including, but not limited to:
In this workshop, we intend to review the most recent developments in the field of data fusion with uncertainty, including, but not limited to:
- Theoretical results in aggregation functions, pre-aggregation functions, and fusion functions with other kinds of weaker monotonicity;
- Theoretical results in aggregation functions, pre-aggregation functions, and fusion functions with other kinds of weaker monotonicity;
- Theoretical results in the controlling of the uncertainty in interval-valued data fusion;
- Theoretical results in the controlling of the uncertainty in interval-valued data fusion;
- Applications in deep learning and adaptative neuro-fuzzy systems, decision making (including, e.g. multi-criteria decision making), image processing, classification and multi-label classification, machine learning, data stream clustering, and data flow prediction.
- Applications in deep learning and adaptative neuro-fuzzy systems, decision making (including, e.g. multi-criteria decision making), image processing, classification and multi-label classification, machine learning, data stream clustering, and data flow prediction.
Organizers
Organizers
Humberto Bustince
Humberto Bustince
Universidad Pública de Navarra
Spain
Javier Fernández
Javier Fernández
Universidad Pública de Navarra
Spain
Cedric Marco-Detchart
Cedric Marco-Detchart
Universidad Pública de Navarra
Spain
Tiago Asmus
Tiago Asmus
Federal Universiy of Rio Grande
Brazil
Hélida Santos
Hélida Santos
Federal Universiy of Rio Grande
Brazil
Workshop Style
Workshop Style
Proceedings:
Proceedings:
The organizers will provide the proceedings by collecting submitted papers. The proceedings will be available online. The proceedings of this workshop will not be included in the proceedings of ECAI 2024.
The organizers will provide the proceedings by collecting submitted papers. The proceedings will be available online. The proceedings of this workshop will not be included in the proceedings of ECAI 2024.
Paper Format:
Paper Format:
Researchers are invited to submit an extended abstract (in PDF) of up to 4 pages (written in English). This extended abstract should include the title, authors, affiliation, a short standard abstract, keywords, the body of the extended abstract, and references. In addition to regular paper submissions, we will consider accepting papers rejected from the main conference, doing so purely based on the reviews written for the main conference.
Researchers are invited to submit an extended abstract (in PDF) of up to 4 pages (written in English). This extended abstract should include the title, authors, affiliation, a short standard abstract, keywords, the body of the extended abstract, and references. In addition to regular paper submissions, we will consider accepting papers rejected from the main conference, doing so purely based on the reviews written for the main conference.
Submissions:
Submissions:
Accepted papers:
Accepted papers:
Constrained Least Squares with Data Fusion via OWA Aggregation Operator (Lucas Almeida; Gildson de Jesus; Eduardo Palmeira)
Edge Detection Ensemble Using Generalizations of the Choquet Integral (Miquéias Amorim Santos Silva; Giancarlo Lucca; Cedric Marco-Detchart; Renan Costa; Eduardo Borges; Graçaliz Dimuro)
Feature-level fusion in wireless acoustic sensor networks with graph attention networks for classification of domestic activities (Wei Wei; Matthias Hutsebaut-Buysse; Thomas Avé; Tom De Schepper; Kevin Mets)
Generalization of Interactive Fuzzy Numbers Concept by Means Aggregation Functions (Eduardo Palmeira; Geizane Silva; José Alfredo Souza)
Introducing the attention mechanism in skip connections of convolutional neural networks (David Erroz; Mikel Galar)
Multi-modal feature extraction combining general knowledge of LLM in advert classification (Tianlin Yang; Christopher Stone)
Restricted dissimilarity functions on arbitrary closed intervals (Tiago Asmus; Giancarlo Lucca; Cedric Marco-Detchart; Humberto Bustince; Heloisa Camargo; Joelson Sartori; Fabiano D'Ávila; Helida Santos; Graçaliz Dimuro)
Towards a Federated Fuzzy Rule Learning Algorithm (Asier Urio-Larrea; Javier Andreu-Perez; Mikel Ferrero-Jaurrieta; Graçaliz Dimuro; Humberto Bustince)
Two-Stage Fine-Tuning for Variable Definition Extraction from Chemical Process-related Papers (Shota Kato; Makishi Yamamoto; Kotaro Nagayama; Manabu Kano)
Page updated
Google Sites
Report abuse