Track: Quality Aspects in Machine Learning, AI and Data Analytics


Machine learning, AI and data analytics have become a major force of research progress in data mining and innovation across enterprises of all sizes. A lot of new platforms with increasingly more features for managing datasets have been proposed in recent years. Given that the datasets are frequently big, data mining is also related to the management of cloud and modern HPC clusters.

Quality assurance in machine learning, AI and data analytics is an important research and engineering challenge in today's data intensive computing. It can be directly related to the quality of data - quality of data generators, privacy, statistical considerations, etc. Given the whole ecosystem surrounding the application of such approaches, quality assurance can also relate to many other aspects, such as the quality of the software implementing the approaches, of the services providing the approaches, of the management of data-intensive computing systems running the approaches, of the relevant resource and data management tools, of the handling of ethical concerns surrounding the use of the approaches, etc. From the machine learning model's point of view, the quality aspects also include model robustness (e.g. generalization, model architecture, resilience to noise), controllability, explainability, and so on.


Papers on this track can explore any topics related to quality in machine learning, AI and data analytics. These include, but not limited to:

  • Quality in data science

  • Quality in deep learning

  • Quality in business intelligence

  • Quality in evolutionary algorithms

  • Quality in fuzzy systems

  • Quality in distributed machine learning systems

  • Data quality in distributed and streaming analytics

  • Algorithms for detecting concept drifts / changes in the underlying distribution of incoming data

  • Algorithms and approaches for detecting outliers, duplicated data, and inconsistent data

  • Efficiency versus accuracy trade-off

  • Data governance

  • Big data quality management

  • Big data quality metrics

  • Big data management across distributed databases and datacentres

  • Big data persistence and preservation

  • Big data quality in cloud systems

  • Testing of machine learning and AI software systems

  • Automated software testing

  • Algorithms and approaches for data healing or system fault healing

  • Procedures for evaluating data models

  • Handling of ethical aspects in data analytics


Chair: Shuo Wang, University of Birmingham, UK

Program Committee:

  • Chun Wai Chiu, University of Birmingham, UK

  • Eduardo Spinosa, Federal University of Paraná, Brazil

  • Honghui Du, University of Leicester, UK

  • Jorge Casillas, University of Granada, Spain

  • João Gama, University of Porto, Portugal

  • Jose Manuel Molina Lopez, Universidad Carlos III de Madrid, Spain

  • Leandro L. Minku, University of Birmingham, UK

  • Tao Chen, Loughborough University,UK

  • Yun Yang, Yunnan University, China

  • Yuwei Guo, Xidian University, China

Shuo Wang is a Lecturer in Machine Learning and Optimization at the School of Computer Science, University of Birmingham (UK). Prior to that, she was a Lecturer in Computer Science at Birmingham City University (UK). She received her PhD degree in Computer Science from the University of Birmingham (UK) in 2011.

Dr. Wang's research interests include data stream learning, class imbalance learning and ensemble learning in machine learning, and their applications in social media analysis, software engineering and fault detection. Her work has been published in internationally renowned journals and conferences, such as IEEE Transactions on Knowledge and Data Engineering and International Joint Conference on Artificial Intelligence (IJCAI).

Among other roles, Dr. Wang was a guest editor of Neurocomputing and Connection Science, and chaired the workshop of IJCAI'17 on Learning in the Presence of Class Imbalance and Concept Drift.