Submit to this session

MLDDA 2025

Special Session on

Machine Learning in Difficult Data Analysis

at the 17th Asian Conference on Intelligent Information and Database Systems (ACIIDS 2025)
23-25 April 2025, Kitakyushu, Japan

Special Session Organizers

Prof. dr hab. Piotr Porwik
University of Silesia in Katowice, Poland
E-mail: piotr.porwik@us.edu.pl

Dr.-Ing Tomasz Orczyk
University of Silesia in Katowice, Poland
E-mail: tomasz.orczyk@us.edu.pl

Dr. Krzysztof Wróbel
University of Silesia in Katowice, Poland
E-mail: krzysztof.wrobel@us.edu.pl

Objectives and topics

Machine learning (ML) is one of the fields of artificial intelligence (AI). ML achievements are currently used in science, business, medicine, biometrics, economics and many other areas of human activity. Machine learning algorithms use various data sets, both labeled and unlabeled. The data may describe two- or multi-class phenomena, and thus may be balanced or may be unbalanced due to the asymmetry of the sample distribution. This type of data should be considered as difficult data. In the first, historic, machine learning methods, these important phenomena were often ignored, which distorted the results. Currently, due to the strong impact of the above-mentioned factors on the quality of automatically made decisions, attempts are made to improve algorithms operating on difficult data. Other factors that make the interpretation of data difficult are the phenomena of seasonality and drift. They involve a random, periodically repeatable or continuous change in the characteristics of the data, affecting their interpretation. All of these are important ML challenges today. The explanation of various interdependencies occurring in the data and their correct interpretation contribute to the creation of new ML algorithms and the discovery of new areas of their applications.
The special session is intended to present experiences resulting from the analysis of difficult data in areas such as:
  • drift detectors,
  • classification of data containing drifts (concept drift, feature drift),
  • classification of seasonal data,
  • interpretation of data in non-Euclidean spaces,
  • completing incomplete data,
  • classification of incomplete data,
  • classification of unbalanced data,
  • neural networks,
  • deep networks,
  • biometrics,
  • biomedicine.

ACIIDS 2025 important dates

Paper submission: January 19, 2025 (hard deadline)
Notification of acceptance: February 10, 2025
Camera-ready papers: February 24, 2025
Registration & payment: February 24, 2025
Conference date: April 23-25, 2025