Special Sessions

Communities around the world are rapidly changing and constantly evolving, with ever new technologies offering great promise for improved health and well-being, safety and security, accessibility and inclusivity, and economic growth. At the same time, the challenges lying at the complex intersection of technology and society have led to a steady increase of interest in highly interdisciplinary approaches that can both benefit from and help advance data science. This special session will bring together researchers, industry experts, practitioners, and potential citizen scientists who are interested in cultivating specialized and important aspects of data science and analytics in the context of smart and connected communities.

Organizers

• Charalampos Chelmis, University at Albany, SUNY, USA
• Ajitesh Srivastava, University of Southern California, USA
• Qi Wang, Northeastern University, USA
• Yasin Yilmaz, University of South Florida, USA
• Daphney-Stavroula Zois, University at Albany, SUNY, USA

Accepted Papers

1. Source detection on networks using spatial temporal graph convolutional networks
   Hao Sha (Indiana University Purdue University Indianapolis), Mohammad Al Hasan (IUPUI), George Mohler

Structures built upon great quantities of networked entities, such as computer networks and social networks, have an undeniable central role in our everyday life. The need to study these complex real-world topologies, together with the growing ability to carry out these studies thanks to technological advances, recently made the use of complex network models pervasive in many disciplines such as computer science, physics, social science, as well as in interdisciplinary research environments.

Organizers

• Martin Atzmueller, Osnabrück Univerity
• Roberto Interdonato, CIRAD, Montpellier
• Rushed Kanawati, University Sorbonne Paris Nord

Accepted Papers

1. A Multilayer Network Perspective on Customer Segmentation Through Cashless Payment Data
   Alessia Galdeman (University of Milan), Cheick Ba (University of Milan), Matteo Zignani (Università degli Studi di Milano), Sabrina Gaito (University of Milan)

Researchers recognize that most of the technical and research challenges that AI experts face when dealing with healthcare resources are data-related. This special session focuses on the challenges of medical data processing and analysis.  The goal of this session is to make the health sector familiar with the ways in which AI can help the health sector and frame data mining research in the context of what medical researchers can expect from their data. Topics addressing the uncertainty of machine learning methods, and explainability of back-box models, interoperability, federated databases, and integration of sources spanning the life of the patient are especially welcome, but we also welcome contributions from the wider domain of medical information processing.

Organizers

• Ernestina Menasalvas Ruiz, Universidad Politécnica de Madrid, Spain
• Alejandro Rodríguez González, Universidad Politécnica de Madrid, Spain

Accepted Papers

1. Towards Treatment Patterns Validation in Lung Cancer Patients
   Guillermo Vigueras (Universidad Politécnica de Madrid), Arturo Redondo (Universidad Politécnica de Madrid), Belén Ríos (Universidad Politécnica de Madrid), Belén Otero (Universidad Politécnica de Madrid), Roberto Hernández (Hospital Universitario Puerta de Hierro), María Torrente (Hospital Universitario Puerta de Hierro), Ernestina Menasalvas (UNIVERSIDAD POLITECNICA DE MADRID), Mariano Provencio (Hospital Universitario Puerta de Hierro), Alejandro Rodriguez (University of Madrid, Spain)
   
   
2. Extracting Cancer Treatments from Clinical Text written in Spanish: A Deep Learning Approach
   OSWALDO SOLARTE (UNIVERSIDAD POLITECNICA DE MADRID), Ernestina Menasalvas (UNIVERSIDAD POLITECNICA DE MADRID), Alejandro Rodriguez (University of Madrid, Spain), MARIA TORRENTE (Hospital Universitario Puerta de Hierro), MARIANO PROVENCIO (Hospital Universitario Puerta de Hierro), ALBERTO BLAZQUEZ (UNIVERSIDAD POLITECNICA DE MADRID)
   
   
3. Interpretable Prediction of Diabetes from Tabular Health Screening Records Using an Attentional Neural Network
   Yuki OBA (University of Tsukuba), Taro TEZUKA (University of Tsukuba), Masaru SANUKI (University of Tsukuba), Yukiko WAGATSUMA (University of Tsukuba)

Environmental and geo-spatial (EnGeo) data is currently obtained by crowdsourcing and public administrations in the context of open data policies. Mining EnGeo data provides relevant insights and potential benefits to public health, medicine, and agriculture.  The analysis of EnGeo data is associated with two major challenges: 1) the integration of heterogenous data; and 2) the selection of the appropriate knowledge discovery process. The main objective of this EnGeoData session is to provide high quality research facing both challenges with theoretical and experimental approaches.

Organizers

• Antonio Lossio-Ventura, National Institutes of Health, USA
• Mathieu Roche, Cirad, TETIS, France
• Maguelonne Teisseire, INRAE, TETIS, France

Accepted Papers

1. 58. Surveillance of airborne plant disease dissemination at continental scale using air mass trajectory analysis and network theory
   Andrea Radici (INRAE)*; Davide Martinetti (INRAE); Daniele Bevacqua (INRAE)
2. 61. Parallel Multi-Graph Convolution Network For  Metro Ridership Prediction
   Fuchen Gao (East China University of Science and Technology)*; Zhanquan Wang (East China University of Science and Technology); Zhenguang Liu (Zhejiang University)
3. 182. On the Use of Class Activation Maps in Remote Sensing: the case of Illegal Landfills
   Rocio Nahime Torres (Politecnico di Milano)*; Piero  Fraternali (Informazione e Bioingegneria); Andrea Biscontini (Politecnico di Milano)
4. 238. A Neural Network Architecture with an Attention-based Layer for Spatial Prediction of Fine Particulate Matter
   Luis E Colchado (Universidad Católica San Pablo )*; Edwin Villanueva Talavera (Pontificia Universidad Católica del Perú); José Eduardo Ochoa Luna (San Pablo Catholic University)

This special session focuses on bringing the research on Explainable Artificial Intelligence (XAI) to actual applications and tools that help to better integrate them as a must-have step in every AI pipeline. We welcome papers that showcase how XAI has been successfully applied in real-world AI-based tasks, helping domain experts understand the results of a model. Moreover, we also encourage the submission of novel techniques to augment and visualize the information contained in the model explanations. Furthermore, we expect a presentation of practical development tools that make it easier for AI practitioners to integrate XAI methods into their daily work.

Organizers

• Victor Rodriguez-Fernandez, Universidad Politécnica de Madrid, Spain
• Szymon Bobek, Jagiellonian University, Poland
• David Camacho, Universidad Politécnica de Madrid, Spain
• Grzegorz J. Nalepa, Jagiellonian University, Poland

Accepted Papers

1. 143. Constructing Global Coherence Representations: Identifying Interpretability and Coherences of Transformer Attention in Time Series Data
   Leonid Schwenke (Osnabrück University)*; Martin Atzmueller (Osnabrück University)    
2. 210. Explainable clustering with multidimensional bounding boxes
   Michał Kuk (AGH University of Science and Technology)*; Szymon Bobek (Jagiellonian University); Grzegorz Nalepa (Jagiellonian University)
3. 236. Explainable artificial intelligence for data science on customer churn
   Carson K. Leung (University of Manitoba)*; Joglas D.N. Souza (University of Manitoba); Adam Pazdor (University of Manitoba)    
4. 244. Explaining Multimodal Errors in Autonomous Vehicles
   Leilani H Gilpin (MIT)*; Vishnu Penubarthi (Massachusetts Institute of Technology); Lalana Kagal (MIT CSAIL)

What do deep learning, chemometrics, graph mining, spatiotemporal data analysis have in common? Tensors!
Tensor analytics are among the most interdisciplinary topics, bringing together a very diverse number of fields and domain applications, with a very successful track record that transcends data science and machine learning. In this special session, we are soliciting original works at the cutting edge of tensor methods for emerging applications, in order to bring together perspectives from the entire spectrum of application domains and methodological advances.

Organizers

• Evangelos E. Papalexakis, University of California Reiverside, USA
• Hadi Fanaee-T, Halmstad University, Sweden

Accepted Papers

1. Unsupervised Tensor based Feature Extraction and Outlier Detection for Multivariate Time Series
   Kiyotaka Matsue (National Institute of Informatics), Mahito Sugiyama (National Institute of Informatics)
   
   
2. A Data-Driven Approach based on Tensor Completion for Replacing "Physical Sensors" with "Virtual Sensors"
   Noorali Raeeji Yaneh Sari (Halmstad University), Hadi Fanaee-T (Halmstad University), Mahmoud Rahat (Halmstad University)

This special session welcomes research papers using Data Mining and Machine Learning (Artificial Intelligence in general) to address the challenges and answer questions related to the problem of predictive maintenance. For example, when to perform maintenance actions, how to estimate components current and future status, which data should be used, what decision support tools should be developed for prognostic, how to improve the estimation accuracy of remaining useful life, and similar. It solicits original work, already completed or in progress.

Organizers

• Bruno Veloso, University of Porto, Porto, Portugal
• Grzegorz J. Nalepa, Jagiellonian University, Krakow, Poland
• Moamar Sayed Mouchaweh, IMT Lille-Douai, Douai, France
• Rita P. Ribeiro, University of Porto, Porto, Portugal
• Sepideh Pashami, RISE, Sweden
• Slawomir Nowaczyk, Halmstad University, Sweden

Accepted Papers

1. Predictive maintenance based on anomaly detection using deep learning for air production unit in the railway industry
   Narjes Davari (INESC TEC), Joao Gama (FEP & INESC TEC), Rita Ribeiro (FCUP & INESC TEC), Pedro Pereira (INESC TEC), Bruno Veloso (Universidade Portucalense & FEP & INESC TEC)
   
   
2. Interpretable Summaries of Black Box Incident Triaging with Subgroup Discovery
   Youcef Remil (Infologic-INSA Lyon), Anes Bendimerad (Infologic), Marc Plantevit (Universit̩é de Lyon, France), Céline Robardet (), Mehdi Kaytoue (Infologic)
   
   
3. Efficient Modeling of Digital Shadows for Production Processes: A Case Study for Quality Prediction in High Pressure Die Casting Processes
   Arnab Chakrabarti (RWTH AACHEN UNIVERSITY), Ravi Prasanna Sukumar (RWTH achen University), Matthias Jarke (RWTH Aachen University), Maximilian Rudack (RWTH Aachen University), Paul Buske (RWTH Aachen University), Carlo Holly (RWTH Aachen University)
   
   
4. Extracting Invariant Features for Predicting State of Health of Batteries in Hybrid Energy Buses
   Mohammed Ghaith Altarabichi (Halmstad University), Yuantao Fan (Halmstad University), Sepideh Pashami (Halmstad University), Peyman Mashhadi (Halmstad University), Slawomir Nowaczyk (Halmstad University)
   
   
5. Explainable anomaly detection for Hot-rolling industrial process
   Jakub Jakubowski (AGH University of Science and Technology,), Przemysław Stanisz (Arcelor Mittal Poland), Szymon Bobek (Jagiellonian University), Grzegorz Nalepa (Jagiellonian University)
   
   
6. Predicting hybrid vehicles' fuel and electric consumption using multitask learning
   Peyman Mashhadi (Halmstad University), Shankara Narayanan Bangalore Ramalingam (Halmstad University), Venkata Sai Vivek Uddagiri (), Mahmoud Rahat (Halmstad University)