CEC 2026 Logo

Tutorial

Evolutionary Feature Reduction for Machine Learning

IEEE Congress on Evolutionary Computation 2026 · June 21–26, 2026 · MECC, Maastricht, The Netherland

Abstract & Learning Objectives

In the era of big data, high-dimensional data are ubiquitous across domains, often leading to the curse of dimensionality and redundancy among features. Feature reduction, through either feature selection or feature construction, aims to produce a smaller, more informative feature set to enhance model efficiency, accuracy, and interpretability. However, the vast search space and complex feature interactions make this task challenging. While exhaustive or traditional heuristic searches are computationally expensive or prone to local optima, evolutionary computation (EC) offers a powerful global search capability for effective feature reduction.

This tutorial introduces the general framework of feature reduction and demonstrates how various EC techniques (e.g., genetic algorithms, particle swarm optimization, differential evolution, genetic programming, ant colony optimization, and evolutionary multi-objective optimization) can be applied to real-world tasks in bioinformatics, image analysis, pattern classification, symbolic regression, and cybersecurity, concluding with open challenges and future research directions.

Outline

  • Introduction to feature reduction
    1. What is feature reduction?
    2. Why is feature reduction necessary?
    3. Illustrating through real-world examples
    4. How to perform feature reduction?
    5. How to categorize feature reduction approaches based on the fitness function?
  • Feature selection: this section reviews existing works based on how feature selection is represented in different EC algorithms (representations)
    1. Graph-based representations:
    2. Tree-based representations:
    3. Vector-based representations: most widely used representations
  • Feature construction:
    1. Why genetic programming for feature construction?
    2. Single-tree representations
    3. Multi-tree representations
  • Hybridisation of feature selection and feature construction: this section reviews an emerging topic in which a subset of original features is combined with a set of new high-level features.
  • Real-world applications of feature reduction: this section illustrates several examples of real-world applications where feature reduction is successfully applied to boost the learning performance.
  • Existing challenges

Length of the tutorial

  • Length of the tutorial: 1.5 hours

Presenters

Biographies

Ruwang Jiao is currently a distinguished professor at Soochow University, China. He currently serves as the Chair of the IEEE Taskforce on Evolutionary Computation for Feature Selection and Construction. He co-chaired the special session at IEEE CEC 2023, IEEE WCCI 2024, and IEEE CEC 2025. He delivered several tutorials at IEEE SSCI and IEEE CEC. He has been invited to serve as a PC member for international conferences such as IJCAI, ECAI, IEEE ICDM, GECCO, IEEE CEC, IEEE SSCI, and Evostar.

Bach Nguyen is a Lecturer in Artificial Intelligence at the CDSAI & School of Engineering and Computer Science, Victoria University of Wellington (VUW). He is the Vice-Chair of the IEEE Task Force on Evolutionary Feature Selection and Construction and the Chair of the IEEE New Zealand Central Section. He is the Publicity Chair of PRICAI 2025. He co-chaired of IEEE Symposium on Computational Intelligence in Data Mining in IEEE SSCI 2021, 2022. He was the organiser of the Special Session on Evolutionary Feature Selection, Construction, and Extraction in IEEE CEC in 2021-2025. He also organized the Special Session on Evolutionary Transfer Learning and Domain Adaptation in SSCI 2021 and 2022. He delivered a Turorial on Evolutionary Feature Reduction in IEEE CEC 2021, IEEE WCCI 2022, IEEE CEC 2023, IEEE WCCI 2024, and IEEE CEC 2025. Dr Nguyen has been serving as a program committee member for over 10 international conferences including AAAI, IJCAI, IEEE CEC, GECCO, and IEEE SSCI.

Bing Xue is a Professor and Deputy Head of School in School of Engineering and Computer Science at VUW. She is also Deputy Director of the CDSAI and Fellow of Engineering of New Zealand. She has over 500 papers published in fully refereed international journals and conferences and her research focuses mainly on evolutionary computation, machine learning, classification, symbolic regression, feature selection, evolving deep neural networks, image analysis, transfer learning, multi-objective machine learning. Prof Xue is a Member of the IEEE Computational Intelligence Society AdCom, a Member of the ACM SIGEVO Executive Committee, and the Vice-Chair of IEEE Task Force on Evolutionary Feature Selection and Construction. She has also served as an Associate Editor of 8 international journal articles including IEEE TEVC, IEEE CIM, and IEEE TAI.

Prof Xue has been organising many international conferences, such as General Chair of PRICAI 2025, Conference Chair of IEEE CEC 2024 and EuroGP 2024. She has also been the organiser of the Special Session on Evolutionary Feature Selection and Construction in IEEE CEC 2015 - 2020. She was a chair for a number of international conferences including the Chair of Women@GECCO 2018 and a co-Chair of the Evolutionary Machine Learning Track for GECCO 2019-2022. She is the Lead Chair of IEEE Symposium on Computational Intelligence in Feature Analysis, Selection, and Learning in Image and Pattern Recognition (FASLIP) at SSCI 2016-2022, a Program Co-Chair of the 7th International Conference on Soft Computing and Pattern Recognition (SoCPaR2015), a Program Chair of the 31st Australasian Joint Conference on Artificial Intelligence, Finance Chair for 2019 IEEE CEC, Tutorial Co-Chair of 2022 IEEE WCCI, and Conference Chair of 2024 IEEE CEC.

高速下载